Date of Award

Spring 5-4-2021

Degree Type


Degree Name

Doctor of Philosophy (PhD)


Political Science

First Advisor

Judd Thornton

Second Advisor

Toby Bolsen

Third Advisor

Sean Richey


The prevailing narrative in political science is that evangelical identity drives the political behavior of evangelicals. This has generated a variety of puzzles for us, in that we struggle to explain how people who are quite similar in terms of their religious beliefs can be quite different in terms of their politics.

I challenge this prevailing narrative. More specifically, I contend that evangelical identity is not the primary determinant of political behavior and that it is outweighed by other relevant factors, such as race and education. Thus, evangelical identity can be more correctly understood as a factor that potentially mitigates the effects of other factors that more directly determine political behavior. In reality, however, this does not occur frequently because most Christians never reach the point in their faith journey at which they truly make their political perspective subject to their faith perspective.

I test this contention using seven measures of political behavior as dependent variables: whether respondents voted for Donald Trump in 2016; a measure of Party Identification; whether respondents support gay marriage; whether respondents support a pro-life position; whether respondents support a ban on assault rifles; whether respondents support legal status for persons brought to the United States illegally as children but who have since graduated from high school in the United States; and whether respondents agree with the statement that racial problems in the United States are rare, isolated situations. The data set is the Common Content section of the 2016 Cooperative Congressional Election Study, supplemented by county level indicators drawn from the 2016 American Community Survey.

The results indicate that, for all seven dependent variables, the strongest effect of evangelical identity is exceeded by the effect of at least one other measure, such as race or education. Further, for all but two of the dependent variables, the strongest evangelical effect is exceeded by the effect of at least one county level indicator, such as the percentage of the county population holding at least a high school diploma. Thus, I conclude that evangelical identity is not the primary determinant of political behavior.


File Upload Confirmation