Civic Religions? The relationship between religiosity and voter turnout in America

By Ariadna Condezo (‘25), Noah DeYoung (‘25), Jacob Kotler (‘24), and David Van Riper (‘26)

Note: This is one of a series of guest posts from students in the Tufts Political Science Research Methods course.

A recent article in The Washington Post reported that “about 1 in 3 Americans claim no religion, a number that […] could grow to more than half by 2070.” Currently, the eventual effects of this stark decline in organized religion are unknown, but one significant area that could be affected is voter turnout. Thus, this project aims to answer: What is the relationship between overall organized religiosity and voter turnout within counties in the United States?

The significance of the study lies in its potential to explain how deeply ingrained social structures, like religious communities, can shape civic behaviors. Investigating this relationship is crucial because religiosity may influence individual values and worldviews, potentially motivating or discouraging voter participation. Understanding these dynamics allows for a better understanding of how religious beliefs and practices are interwoven with political behavior, offering insights about how the current decline in American religion may affect civic engagement and political participation.

We hypothesized that overall religiosity will show a positive correlation with voter turnout, as religious communities often promote voting through discussions of civic duty and community engagement.

Data

To investigate the relationship between religiosity, our primary independent variable, and voter turnout, our dependent variable, we combined multiple county-level datasets. Our measure of religiosity is from the U.S. Religion Census, which is a decennial survey conducted by the Association of Statisticians of American Religious Bodies. We used the adherents as a percent of population variable from the 2020 census. Our measure of voter turnout is from the National Neighborhood Data Archive (NaNDA).We used the voter turnout variable from 2018, which is the percent of eligible voters who cast a ballot in the election. We also used a partisanship index variable as a control in our regression analysis.

We used a third dataset County-level Socioeconomic Data for Predictive Modeling of Epidemiological Effects, which was compiled by students and faculty at John Hopkins University to add our other control variables: population density, median income, percent of population over 65 years old, and percent of population with at least a bachelor’s degree. This dataset used multiple sources, but primarily the US Census.

We excluded all counties in the state of Wisconsin, Rockwall County, and Piatt County due to anomalies in their data entries. We also excluded counties with a population of less than 5,000 people to reduce noise in our data. After cleaning our data, we had 2,692 of 3,143 counties remaining in our dataset.

What we found

The results of our regression analysis can be seen in the figure below. Overall, we found that a county’s level of organized religiosity has a positive relationship with its level of voter turnout. In our model, organized religiosity had a coefficient of 0.018, meaning that for every 1 percentage point increase in a county’s level of organized religiosity, its voter turnout rate increased by roughly 0.02 points. This result was statistically significant at the 95% confidence level, meaning we can be 95% confident that there exists a positive relationship between organized religiosity and voter turnout on the county level. However, despite this high level of statistical significance, these results do not appear to be substantively significant, as the size of the effect is incredibly small. For example, for a county’s voter turnout rate to increase by even a single percentage point, its level of organized religiosity would have to increase by 50 percentage points. Therefore, while we the relationship between organized religiosity and voter turnout appears to be a positive one, this relationship is relatively inconsequential.

Note: Bars represent 95% confidence levels. While not shown in this plot, this model also controls for state.

Beyond our main independent variable of organized religiosity, our model also produced various other statistically significant results. Perhaps the most notable of these is that we found that the percentage of a county’s population that is above age 65 has a very strong positive relationship with its voter turnout rate. Not only is this result statistically significant at the 95% confidence level, but this result is also substantively significant. In fact, we found that for every 1% increase in a county’s 65+ population, its voter turnout rate increased by 0.86%, a near 1–1 ratio. In addition, contrary to common perception, our model also found that the percentage of a county’s votes that were for Democratic candidates had a statistically significant positive relationship with voter turnout. However, it is important to note that our analysis focuses on the 2018 elections, which was a cycle framed as a national rebuke of Trump and the GOP as a whole. Thus, it is perhaps not surprising that we found that, in this particular year, the more Democratic a county was, the higher its voter turnout was as well.

Conclusion

Our project yielded several notable findings. A county’s level of voter turnout has a positive relationship with its level of organized religiosity. However, the small magnitude of our results suggests that organized religiosity plays a relatively minor role in driving voter turnout in U.S. counties. Further studies could explore the nuances of how different denominations within organized religiosity affect voter turnout. Understanding these dynamics could reveal more significant effects within specific religious communities. Certain denominations or religious traditions may place more value on activities like voting as an expression of responsible citizenship than others. Different faiths’ emphasis on moral principles may also influence how important political participation is to their followers. Still, based on these results, it appears that the decline in American religiosity will not necessarily have a major effect on voter turnout.