The Role of Religious Congregations in the AIDS Crisis of Sub-Saharan Africa Jenny Trinitapoli Dissertation, Department of Sociology, The University of Texas at Austin ABSTRACT The primary goal of this dissertation is to examine how religious organizations (ROs) and their “moral communities” influence responses to the AIDS epidemic in rural Malawi, a setting with both high levels of HIV prevalence and high levels of religious participation. Recent studies connecting religion and HIV risk behaviors have uncovered new support for an old, Durkheimian idea – that participation in harmful behaviors is reduced in places where particular religions or religious rituals are widely practiced. This “moral communities” thesis provides a helpful correction to the Western tendency to focus only on individual traits and behaviors, a focus that seems misguided in sub-Saharan Africa as well. This dissertation assesses the roles of ROs in both HIV prevention and AIDS mitigation using multiple data sources and mixed methods. One of the central assumptions of early sociological research was that a proper understanding of human behavior included the joint consideration of individual religiousness and the social/religious contexts in which those traits have meaning. In this dissertation, I conceptualize congregations as “moral communities” that influence both individual attitudes and behavior relevant to prevention as well as perceived obligations to support families affected by AIDS. This dissertation is organized into three parts, each focusing on the relationship between religion and a distinct dimension of the HIV crisis: HIV risk behaviors, HIV status, and HIV-related stigma. Part 1 examines three outcomes that correspond with the infamous ABCs of HIV prevention: abstinence (for unmarried individuals), fidelity (for married persons), and condom use, testing the relationship between the characteristics of religious organizations (doctrines, strictness) and the reported behavior of individuals. Consistent with previous research, high levels of religious involvement are associated with greater likelihood of reporting abstinence and faithfulness, however highly religious individuals are less likely to report using condoms, which may put them at increased risk if/when they do engage in risky sexual practices. Multi-level models of congregational effects show that while religious messages about sexual morality in general do not influence reports of sexual behavior, direct discussion of AIDS in weekly religious services does. Adolescents attending congregations in which the leader reports frequently discussing AIDS in weekly religious services are significantly less likely to report being sexually active than those who attend congregations where AIDS is not discussed at all or is discussed only occasionally. While messages given in religious services do not have an impact on self-reported condom use, the views and advice of religious leaders do. Individuals belonging to congregations in which the leader reports privately advising members to use condoms are more likely to report using them. A truly mixed methods endeavor, this chapter also utilizes indepth interviews with religious leaders to document patterns and variations in the formal and informal messages about AIDS and sexual morality individuals likely hear when they attend their churches and mosques and uses these to explain some of the patterns observed in the quantitative models. While most of the studies on religion and HIV rely on self reports of risk behavior, Part 2 of my dissertation looks directly at the association between religious involvement and actual HIV

status, using unique biomarker data available from wave 3 of the Malawi Diffusion and Ideational Change Project (MDICP-3). This analysis pays special attention to five possible mechanisms that might produce this relationship: 1) religious affiliation; 2) religious involvement; 3) selectivity; 4) differential risk behaviors; 5) contextual religious effects. At the individual level, religious involvement is strongly associated with HIV status for both men and for women. Issues of selectivity (health selectivity, in particular) partially account for this relationship, but do not explain it completely. Surprisingly, while religious involvement is associated with key sexual risk behaviors, these behaviors do not predict actual HIV status, posing an interesting dilemma for understanding how religious involvement may be affecting HIV status. Multi-level models examining the role of village context offer a partial explanation for these findings. These models reveal that community-level factors, including the religious context of a village, are far superior predictors of HIV status when compared with the individuallevel characteristics traditionally employed in studies of HIV risk or infection. Finally, Part 3 shifts the focus to the question of how religious organizations are responding to AIDS in their villages. This chapter examines how religious organizations both promote and combat AIDS-related stigma and discrimination by examining AIDS-related attitudes and behaviors at the institutional level. In contrast to the “variable-centered” analytical approach used in the other portions of this dissertation, this portion of the analysis takes more of a grounded theory approach - focusing on the systematically selected cases of congregations at high and low ends of a discrimination scale, and documenting the emergent themes that will help develop a better understanding of the ways congregations can both facilitate and impede effective responses to the AIDS epidemic in a HIV-prevalence rural setting. This dissertation provides the first systematic and empirically-based assessment of how religious organizations influence responses to the AIDS epidemic in a high prevalence region of subSaharan Africa. Using multiple sources of data and multiple methods, this study examines three distinct aspects of the HIV crisis: risk behavior, seroprevalence, and AIDS-related stigma and discrimination. In doing so, it contributes to developing more complex understandings of the contours of AIDS epidemic in this region, as well the factors that facilitate and impede effective and responses to it. In addition, this dissertation makes contributions to the sociology of religion overall – in particular to the field of religion and health.