The Politics of Disaffiliation: How Religious Disaffiliation Impacts the Partisan Preferences of Former Christians, Existing Christians, and the Non-Religious

The United States is experiencing rapid change in its religious demography. While the U.S. has historically maintained a sizable Christian majority, the share of Americans identifying as Christian is expected to fall below fifty percent by 2060. This change corresponds with sharp growth in the share of Americans who are non-religious. These trends are deeply related—being primarily driven by individuals who were raised Christian but no longer identify as Christian later in life. About six in ten non-religious Americans—one in five U.S. adults—are non-religious former Christians. Despite the magnitude of these religious demographic trends, there has been insufficient scholarly attention examining how they may affect American politics.

The purpose of this dissertation is to explore the politics of disaffiliation—focusing specifically on how the growing share of Americans who have disaffiliated from Christianity is impacting support for the two major political parties. This is done in two stages. The first stage examines the within-subject effect of disaffiliation on the partisan preferences of former Christians; the second stage examines how making religious demographic trends salient impacts the partisan preferences of existing Christian and non-religious Americans.
In brief, I find that, contrary to the popular assumption that these demographic trends will serve as political tailwinds for the Democratic Party, due to the distinctively higher rates of Democratic support among non-religious Americans compared to Christians, these religious demographic trends do not seem to decisively benefit either major political party.

Specifically, in Stage One, I draw on both “religion-first” and “politics-first” frameworks of understanding religion and American politics and find that disaffiliation from Christianity and becoming non-religious is only associated with a small within-subject increase in Democratic support that is confined to White former Christians—attributable to the adoption of more liberal values and political attitudes. This shift is in part counteracted by the segment of Black former Christians leaving mostly Black churches, who become more likely to support Republicans after disaffiliation, which I argue is due to no longer experiencing racialized social pressure to support the Democratic Party in church.

In Stage Two, I draw on the fast-growing literature on White Christian nationalism and find support for the view that making disaffiliation salient increases Republican support among exclusively White Christians by evoking racial and religious social status threat among this group. Conversely, while Black and Latino Christians experience religious status threat when reminded of religious demographic trends, I find no evidence that the partisan preferences of these groups are impacted.

Finally, building upon research showing that descriptive representation is positively associated with political efficacy among minority groups, I find suggestive evidence that making religious demographic trends salient corresponds with increases in both religious group political efficacy—i.e., a novel measure of confidence in the government’s responsiveness to the collective political needs of one’s religious group—and partisan-ideological sorting (i.e., liberals are more likely to support Democrats and conservatives are more likely to support Republicans), which does not clearly benefit either major party despite non-religious Americans skewing more liberal than conservative.
Viewed collectively, the fact that these trends do not appear to provide a significant net advantage to either major party does not mean they will have a negligible impact on American politics. One reoccurring observation throughout this dissertation is that religious demographic trends are contributing to higher levels of social sorting across many different social groups. This matters because existing research indicates that higher levels of social sorting can have undesirable consequences, including increased affective polarization, greater distrust of out-groups, heightened political anger, declining respect for democratic norms, and intensified political gridlock. Given mounting concerns that worsening is fueling support for political violence and leading to the erosion of democratic institutions, the findings presented in this dissertation suggest that religious demographic change is serving to widen the very divisions threatening the stability of American democracy itself.