Abstract
We interpret three strands of social movement theory (social threat, opportunity structure, and resource mobilization) to understand religious congregations' political activity related to same-sex marriage. Using a unique, national data set from the 2008 presidential election, we show the importance of the anti-same-sex marriage discursive opportunity structure (DOS) for facilitating, and occasionally constraining, congregational political activity. Other theoretical factors influence congregations under limited, although important, conditions. Our research develops a nuanced understanding of congregations' roles in social movements that prioritize institutionalized political activity and expands DOS theory in relation to local religious organizations.
ACKNOWLEDGMENTS
The authors wish to thank the following for helpful comments: Kraig Beyerlein, Melinda Kane, and the participants at the 2010 ASA Religion Section Session on Religion and Movements for Equality and Human Rights. We thank Corwin Smidt for access to the data set.
NOTES
Notes
1 For ease of usage, we sometimes employ the phrases anti-gay and pro-gay, but include anti-lesbian and pro-lesbian in our meaning.
2 Here, we mean SMOs with state targets, not intra-organizational social movements with nonstate targets (CitationZald 1982).
3 No cases from the state of Vermont reported monthly religious attendance or more.
4 Survey question rotation affects the distribution of three variables, underrepresenting congregations in a ballot state and counties with black residents, while overrepresenting non-Christian congregations (p < 0.05).
5 Since this question came at the end of a series of questions about the specific congregational context, we assume that the respondent's focus was on discussion within the congregation. The use of “discuss” implies an interaction or series of interactions that would most likely occur at the local congregational level.
6 The average count of congregations per state is 8.3, with only 7 (15 percent) of 48 states containing singletons.
7 We thank a reviewer for this suggestion.
8 Given the importance of weighting such descriptive measures when using hypernetwork sampling, and our lack of weights despite organizational size information, we present these frequencies only to describe the data.
9 and were constructed using the Stata 12 margins command. After estimation by xtlogit, the group intercept was fixed at the mean and the marginal effects were averaged over other explanatory variables to show the nonlinear probability of the outcome.
10 was constructed using the Stata 12 margins command. For ease of viewing this conditional effect, after estimation by xtlogit, the figure shows the linear prediction of the outcome, with the group intercept fixed at the mean and marginal effects averaged over other explanatory variables.