Attitudes toward homosexuality in Latin America and Chile

There is abundant research on the increased acceptance of homosexuality all over the world since the end of World War II. People have become more open minded and respectful of diversity, including homosexual relationships (Gibson Reference Gibson1992; Zhang and Brym Reference Zhang and Brym2019). Nonetheless, according to Roberts (Reference Roberts2019, 42), this shift has not been homogeneous.

Regarding trends over time, there is evidence of a broad global upswing in the acceptance of homosexuality between 1981 and 2012. Change has not only occurred in Western countries. Generally speaking, global attitudes toward homosexuality also became more positive over this time, albeit varying considerably from one country to another.

Although a global analysis shows a general decrease in homophobic attitudes, a closer look at the evidence reveals that this reduction has been more intense in Western societies, and numerous countries elsewhere continue recording high levels of intolerance toward homosexual relationships (Jäckle and Wenzelburger Reference Jäckle and Wenzelburger2015, 208). In this sense, a 2019 Pew Research Center survey shows that between 2002 and 2019 there was a general worldwide increase in the acceptance of homosexuality, especially in South Africa, India, Turkey, and Japan. However, this research also revealed that some countries have not recorded any statistically significant progress (Poushter and Kent Reference Poushter and Kent2020, 25).

Latin America has historically been a hostile environment for individuals who identify as lesbian, gay, bisexual, transgender, intersex, queer, asexual, or otherwise through several discriminating attitudes, including violence (e.g., murder, aggression, sexual abuse) and social and labor market exclusion (Encarnación Reference Encarnación2011).

Over the past decades, though, most countries in Latin America have enacted far-reaching legislative reforms designed to foster acceptance of homosexuality and, in particular, to decriminalize same-sex relationships. Remarkable progress has been made in the region in terms of the acceptance of same-sex relationships (Roberts Reference Roberts2019), albeit with widely differing levels, with major differences between countries such as Argentina and Uruguay and others such as Guatemala and Nicaragua (Lodola and Corral Reference Lodola and Corral2010). According to the Latinobarometer (Corporación Latinobarómetro 1998, 2009), the percentage of people who would not like to have LGBTQIA+-identifying individuals as neighbors fell from 51 percent to 29 percent in Latin America between 1998 and 2009. Chile recorded a similar trend, with a decrease from 44 percent to 23 percent over the same period. ^{Footnote 1}

Chile’s attitude toward homosexuality was highly influenced by the Pinochet dictatorship, which promoted conservative values and condemned any kind of nontraditional sexuality (Carvajal Reference Carvajal2019). This hostility continued even after the restoration of democracy in 1990, thus leading to a relative delay—compared to other Latin American countries with higher levels of human development—in the implementation of policies recognizing sexual rights. For example, in 1999 Chile was the penultimate country in Latin America to decriminalize same-sex relationships. Thirteen years later, in 2012, following the murder of a gay teen, Daniel Zamudio, by a neo-Nazi group, Chile passed an antidiscrimination law that included protections for the LGBTIQA+ community. Nonetheless, this approval required seven years of debates in the Chilean congress and faced strong opposition from conservative civil and religious organizations (Díez Reference Díez2015). The Civil Union Bill (Acuerdo de Unión Civil), which includes same-sex relationships, was not passed until 2015. A year later, article 2 of the 2016 Modernization of Labor Relations Act No. 20,940 reformed the Chilean Labor Code, explicitly forbidding any kind of employment discrimination because of sexual orientation (Mendos Reference Mendos2019).

In 2016, Chile was one of the Organisation for the Co-operation and Economic Development (OECD) countries with a worst record in terms of recognition of LGTBIQA+ rights. By then, Chile scored only six out ten points in the OECD’s list of items for assessing the legal recognition of homosexual orientation (Valfort Reference Valfort2017). However, in the past years, Chile has experienced a remarkable evolution toward the recognition of homosexual rights. It legalized same-sex marriage and adoption in December 2021 (a reform that came into force in March 2022). ^{Footnote 2} Moreover, the draft of the constitution rejected in the plebiscite in September 2022 aimed to prohibit discrimination based on sexual orientation (already forbidden by the criminal law). Even though age limits are still different for heterosexual and homosexual acts (fourteen and eighteen years old, respectively), a proposal to reform the Criminal Code is currently under discussion. ^{Footnote 3}

All in all, since 1990 Chileans have experienced a constant increase in the acceptance of sexual diversity according to different data sources (Barrientos Reference Barrientos2015). The reasons for the general time lag in all these changes—compared to Argentina or Uruguay, for example—are still being debated (Contardo Reference Contardo2011; Díez Reference Díez2015; Corrales Reference Corrales2017). Schulenberg (Reference Schulenberg2019, 13–14) has pointed to the fragmentation and weakness of the LGBTQIA+ movement, historically divided into various factions, including the Movimiento de Integración y Liberación Homosexual (Movement for Homosexual Integration and Liberation), which has split internally into the Movimiento Unificado de Minorías Sexuales (Unified Movement of Sexual Minorities) and Fundación Iguales (Equals Foundation).

Theoretical framework and literature review

There is a large corpus of literature that seeks to identify variables that may shape attitudes toward homosexuality. This research also suggests the importance of both demographic characteristics and country-level factors.

There are numerous socioeconomic drivers that can potentially explain the formation of attitudes toward homosexual relationships. First, males are often less accepting of homosexuality than women (I-Huston and Waite Reference I-Huston and Waite1999). In this respect, several studies have posited that men have stricter expectations of masculinity, so tend to reject homosexual males more (Kite and Whitley Reference Kite and Bernard1996; Louderback and Whitley Reference Louderback and Bernard1997).

This literature also identifies the relevant role of location of residence: it offers consistent evidence that residents in rural areas are less accepting of homosexual relationships than the urban population is. Some researchers have emphasized the importance of community size: while those individuals living with people similar to themselves (“provincials”) tend to develop individual or communitarian attitudes, those living in larger urban locations (“cosmopolitans”) are more likely to be in contact with a more heterogeneous community, thus developing a more open-minded attitude (Andersen and Yaish Reference Andersen and Yaish2003; Andersen and Fetner Reference Andersen and Fetner2008a).

Complementarily, birth cohort is a potential factor for analyzing acceptance of same-sex relations. In this sense, exposure to certain social, historical, and cultural events can have an impact on the development of individual attitudes. Therefore, the sociocultural environment surrounding individuals can consolidate social generations joined by particular views or perceptions (Haavio-Mannila, Roos, and Kontula Reference Haavio-Mannila, Roos and Kontula1996, 410). In this sense, more recent birth cohorts are more likely to show more tolerant attitudes toward sexual diversity, including same-sex relations (Treas Reference Treas2002; Andersen and Fetner Reference Andersen and Fetner2008a; Lewis and Gossett Reference Lewis and Gossett2008; Kranjac and Wagmiller Reference Kranjac and Wagmiller2021).

According to the existing literature, age is another driver of these kinds of attitudes, as it is often positively related to socially conservative attitudes, including nonacceptance of same-sex relationships. Younger people are usually more accepting of homosexuality than older people (Quillian Reference Quillian1996; Andersen and Fetner Reference Andersen and Fetner2008a; Hooghe and Meeusen Reference Hooghe and Meeusen2013; Navarro et al. Reference Navarro, Barrientos, Gómez and Bahamondes2019; Seligson, Moreno Morales, and Russo Reference Seligson, Moreno Morales and Russo2019). According to Adamczyk and Pitt (Reference Adamczyk and Pitt2009), the relationship between nonacceptance of same-sex relationships and age is especially intense in contexts of political and economic instability.

It is worth mentioning the existence of an ongoing discussion on the modeling of the age cohort and period to understand different generational processes (Bell and Jones Reference Bell and Jones2014). Attitudes like acceptance of homosexuality can be a product of age differences, birth cohort effects, period effects, or a combination of several factors. In this sense, while there are life events that might affect the formation of beliefs about certain topics, there is also a broad consensus that attitudes toward controversial issues stem from generational differences, and there are rarely major discrepancies within each birth cohort (Andersen and Fetner Reference Andersen and Fetner2008a).

The importance of education is also a constant in the studies seeking to understand the main determinants of homophobic beliefs and behaviors. It is generally agreed that higher levels of education are related to higher levels of acceptance of homosexuality (Pampel Reference Pampel2016; La Roi and Mandemakers Reference La Roi and Mandemakers2018), with some researchers attributing this behaviour to the “educational effects explanations,” that is, schooling provides cognitive and reasoning attitudes that involve the assimilation of new views (Campbell and Horowitz Reference Campbell and Horowitz2016). A second branch of literature suggests that liberal values—including acceptance of sexual diversity—are socialized as part of higher education plans (Carvacho et al. Reference Carvacho, Andreas Zick, Roberto González, Kocik and Bertl2013; Slootmaeckers and Lievens Reference Slootmaeckers and Lievens2014). Third, the effect of education might be due to contact with other students and the informal interaction between students and teachers encouraging openness to different views (Pettigrew and Tropp Reference Pettigrew and Tropp2006; Andersen and Fetner Reference Andersen and Fetner2008a). Furthermore, some studies suggest that the role of education is stronger in more developed countries (Seligson, Moreno Morales, and Russo Reference Seligson, Moreno Morales and Russo2019), while its role is limited in authoritarian regimes, which may use education as a means of indoctrination (Zhang and Brym Reference Zhang and Brym2019).

There is compelling evidence to show that higher levels of religiosity in an individual are related to a greater intolerance of homosexuality, both in terms of self-identification (belonging to a religion) and attendance (participating in the religious community) (Rowatt et al. Reference Rowatt, Jo-Ann Tsang, LaMartina, McCullers and McKinley2006; Adamczyk and Pitt Reference Adamczyk and Pitt2009; Adamczyk Reference Adamczyk2017). According to Janssen and Scheepers (Reference Janssen and Scheepers2018, 16), “it seems that every dimension of religiosity has a positive relationship with rejection of homosexuality.” In addition, there is also evidence to show that the intensity of this nonacceptance is highly determined by an individual’s specific religion: while Hinduism, Protestant Evangelicalism, and Islam usually involve higher levels of rejecting homosexual unions, Judaism and Buddhism have higher levels of acceptance (Jäckle and Wenzelburger Reference Jäckle and Wenzelburger2015; Navarro et al. Reference Navarro, Barrientos, Gómez and Bahamondes2019).

Contemporary research also links an individual’s ideological orientation to acceptance of homosexuality. In particular, conservatives tend to record higher levels of rejecting same-sex relationships than liberals (Barth and Parry Reference Barth and Parry2009; Sherkat et al. Reference Sherkat, Powell-Williams, Maddox and Mattias De Vries2011). Jost (Reference Jost2006) links this stance to a reluctance to change and opposition to more egalitarian politics. Van der Toorn and colleagues (Reference van der Toorn, Jost, Packer, Noorbaloochi and Van Bavel2017) suggest that this attitude among conservatives is closely correlated to religiosity and sexual prejudice.

Social capital has also received considerable attention from researchers. According to Putnam (Reference Putnam1995, 35), social capital is understood to be the “features of social organizations, such as networks, norms and trust that facilitate action and cooperation for mutual benefit.” Accordingly, the more social capital a community has, the stronger the ties among its members and the more legitimated its authorities are. Concerning acceptance of homosexuality, authors such as Rohe (Putnam et al. Reference Putnam, Light, Briggs, Rohe, Vidal, Hutchinson, Gress and Woolcock2004) and Walters (Reference Walters2002) highlight the role of civic engagement and social embeddedness in explaining attitudes toward homosexuality. There seems to be a general consensus on how higher levels of social trust and connection to others are related to higher levels of acceptance, including those of homosexual relationships (Lee Reference Lee2014; Jones Reference Jones2015). Nevertheless, despite the overall link between participation in associations and acceptance, belonging to certain specific groups—such as veterans’ associations or church congregations—might have a negative impact on attitudes toward homosexuality. Specifically, this research adopts a twin approach to measure social capital. On the one hand, we analyze trust in people, which Putnam (Reference Putnam1995, 230) has signaled as crucial for understanding social connectedness and civic engagement; on the other hand, we use trust in institutions as an intermediate variable that reflects the overall level of social capital. These two variables are used to integrate both the informal and formal aspects of social capital, respectively (Nooteboom Reference Nooteboom2007).

The results of the specific research on Latin American and Caribbean countries are consistent with the findings of international literature, highlighting the role of gender, religion, education, ideology, location of residence, economic and social development, and education as the main predictors of acceptance of homosexuality (Lodola and Corral Reference Lodola and Corral2010; Chaux and León Reference Chaux and León2016; Navarro et al. Reference Navarro, Barrientos, Gómez and Bahamondes2019; Seligson, Moreno Morales, and Ruso Reference Seligson, Moreno Morales and Russo2019). There is also some literature focusing solely on the Chilean case, such as studies by Cárdenas and colleagues (Reference Cárdenas, Barrientos, Gómez and Frías-Navarro2012) and Cárdenas, Barrientos, and Gómez (Reference Cárdenas, Barrientos and Gómez2018). By analyzing relatively small samples of university students (fewer than three hundred individuals), these authors highlight the role of gender, religiosity, gender role beliefs, and authoritarian attitudes in line with the findings mentioned here.

Data

This study uses the second and fourth waves of the ISSP Religion Module for Chile conducted in 1998 and 2018 (ISSP Research Group 2000, 2020). These cross-sectional surveys in Chile are administered by the Center of Public Studies, a private nonpartisan and nonprofit academic foundation of renowned experience and reputation in such studies. This institution uses a three-stage sampling and region-level stratified design. This module collects information on religious attitudes and beliefs together with a relatively rich set of socioeconomic characteristics of the Chilean population eighteen years of age or older. The initial sample sizes of the 1998 and 2018 waves are 1,500 and 1,400 observations, respectively. ^{Footnote 4}

The key feature of this survey is that it captures attitudes toward homosexuality in a homogeneous way through the following question: “Do you think it is wrong or not wrong if two adults of the same sex have sexual relations?” The possible answers include “always wrong,” “almost always wrong,” “wrong only sometimes,” “not wrong at all,” and “can’t choose, don’t know.” To make the econometric analysis more tractable, we recode the variable into two categories: same-sex relationships are not wrong at all according to the respondent and other possibilities. Our derived variable takes the value of 1 if the individual does not consider this type of relation to be wrong and 0 otherwise.

Furthermore, based on the previous literature on attitudes toward homosexuality, we consider the following potential explanatory variables of interest: gender, location of residence (rural or urban), birth cohort, educational attainment, religiosity, political affiliation, trust in other people, and trust in institutions (as the two dimensions of social capital). As mentioned earlier, the role of age in shaping acceptance of homosexuality is not totally clear (Andersen and Fetner Reference Andersen and Fetner2008a). Its role might be just a result of birth cohort effects, period effects, or a combination of the two. The methodology used here, outlined below, requires the inclusion of period effects. Our baseline estimation considers cohort effects, and we present the results including age effects in the supplementary online material (Tables S2 and S3).

Given that the sample sizes are moderate, we recode some variables to avoid multicollinearity issues. First, we consider individuals to be religious if they declare themselves to be (irrespective of the strength of this commitment). Second, we look at political affiliation as a proxy for ideology, which is not available in this survey, considering the following four categories: left (voter of any of the parties that makes up the Concertación, or the New Majority, apart from the Christian Democratic Party, and other more leftist political groups), center, right (the National Renewal Party and the Independent Democratic Union), and last, any other affiliation (all other possibilities, including abstention or blank voting). Third, following the discussion in the previous section, we consider two measures of social capital. The first uses the respondent’s degree of trust in people, a binary variable computed from a question with four possible levels of trust. This variable is recoded as 1 if the individual considers that people can always or usually be trusted and as 0 otherwise. The second one draws on a battery of questions on trust in different national institutions (parliament, church and religious organizations, courts and the legal system, and school and the education system). We create a binary variable with a value of 0 or 1 for each institution covered by these questions (based on whether the respondent expressed trust in them). The answers to the four questions are summed to calculate an index of confidence in institutions ranging from 0 (no confidence in any of the institutions) to 4 (confidence in all four institutions covered).

The rationale for the inclusion of the variables as covariates comes from the findings highlighted in prior literature. First, women tend to be more positive toward homosexuality than men. Second, the view of people living in urban areas usually evidences higher levels of acceptance of homosexual relationships than their rural counterparts. Third, these attitudes are expected to become more negative with age. Fourth, education should have a positive effect on the acceptance of homosexuality. Fifth, conversely, the prior literature suggests that religiosity might have a detrimental impact in this matter. Sixth, according to previous evidence, we expect right-wing individuals to be less prone to agree with homosexual relationships. Seventh, the higher the level of social capital, the more positive is the expected attitude toward homosexuality.

Methods

The main descriptive statistics are presented for the evolution of attitudes toward homosexual relationships in Chile between 1998 and 2018. Our main task consists in identifying the factors that drive the changes in the acceptance of homosexuality among the Chilean population over time. An econometric decomposition is performed to disentangle those changes due to individual socioeconomic characteristics and those attributed to structural changes related to how the views of people with certain demographic features evolve over time.

Our approach consists of applying a version of the well-known Oaxaca-Blinder decomposition (Blinder Reference Blinder1973; Oaxaca Reference Oaxaca1973) adapted to binary variables, closely following the procedure proposed by Gomulka and Stern (Reference Gomulka and Stern1990) and Yun (Reference Yun2004) (for some applications of this decomposition, see Gang, Sen, and Yun Reference Gang, Sen and Yun2008; Muñoz de Bustillo and Antón Reference Muñoz de Bustillo and Antón2010; Antón, Muñoz de Bustillo, and Fernández Macías Reference Antón, Muñoz de Bustillo and Fernández-Macías2014).

Specifically, the method unfolds as follows. First, we separately estimate logit models for 1998 and 2008 to explore the determinants of the acceptance of homosexuality in those years. We record the changes over time according to the variation due to the differences in characteristics by adding and subtracting the predicted probability of accepting homosexuality if the population in 1998 determined their attitudes as individuals interviewed in 2018. This means separating the twenty-year difference in a component due to the change in characteristics, and therefore capturing the differences existing if the population in 1998 behaved as respondents in 2018, and the discrepancy associated with structural changes in the responses (i.e., the differences related to the changes in the coefficients, obtained by applying the coefficients of both years to the population in 2018). The covariates included in the analysis are those discussed in the previous section. The explained and unexplained parts corresponding to each variable are computed by following the solution proposed by Yun (Reference Yun2005, Reference Yun2008). ^{Footnote 5}