Religion and the Rainbow Struggle: Does Religion Factor Into Attitudes Toward Homosexuality and Same-Sex Civil Unions in Brazil?

Abstract

The provision of civil liberties to LGBT persons has become part of a global movement in societies across the world. In Brazil, a recent judicial ruling for the first time established the right for homosexual couples to enter into civil unions, despite the presence of widespread disapproval of homosexuality among the population and opposition from prominent religious groups. Picking up on this issue, the following study examines whether religion may factor into the attitudes Brazilians hold toward homosexuality and same-sex civil unions. Using data from the Brazilian Social Research Survey, we find that the most restrictive views toward homosexuality and the strongest opposition to same-sex civil unions are most prevalent among devoted followers of historical Protestant, Pentecostal, and Catholic faith traditions, whereas adherents of Afro-Brazilian and spiritist religions, as well as those with no religious commitment, are inclined to assume a more tolerant moral posture toward such issues. The findings point to religion as a potential influence in future public policy initiatives and social movements involving LGBT issues in Brazil.

KEYWORDS:

• Brazil
• religion
• Protestantism
• homosexuality
• same-sex marriage
• Latin America

Notes

1. For more information on Kardecist Spiritism, see Kardec (2006).

2. The Brazilian Social Research Survey (PESB: Pesquisa Social Brasileira) was coordinated by Alberto Carlos Almeida, Andréia Schroeder, and Zairo Cheibub from the Universidade Federal Fluminense with a grant from the Ford Foundation. Data and complete details about survey methodology for the BSRS are available at http://www.nadd.prp.usp.br/cis/index.aspx. The PESB is the most recently available data set containing the variables of focal interest to this study.

3. We dichotomized both these variables because of a heavily skewed distribution on the original ordinal measures, which created problems for employing the ordered-logit regression model. We also considered creating a single, continuous index measure of both homosexuality variables, but the frequency distribution rendered unacceptable skewness and kurtosis values for the OLS regression model.

4. This variable does not intend to measure any specific legislation or legislative proposal with regard to same-sex unions. It is important to note that the BSRS was administered when civil unions in the country were unrecognized by law.

5. The category of Spiritism denotes Afro-Brazilian and spiritist faith traditions and includes followers of Umbanda, Candomble, and Kardecist Spiritism. Although these three groups represent distinct faith traditions in Brazil’s religious landscape, we decided to collapse these three groups into one category (i.e., Spiritism) for our analysis. A bivariate analysis revealed a similar statistical correlation between followers of each of these groups and attitudes toward homosexuality and same-sex unions. The cases were not sufficient on their own for each group to justify an individual coding category for the multivariate analysis. Thus, rather than exclude these groups from the study, we opted to create a unique category that encompasses all three.

6. The variables religion, educational attainment, marital status, and region of residence are coded as “dummy” variables.

7. These findings are reported in the Results section but are not tabulated (full tables available upon request).

8. The BSRS was carried out in 2002. It is the most current survey data that contains the variables of interest for this research.