2024 Theses Doctoral

Education For Democracy: Three Essays Exploring the Relationship Between Education and Voting

Arnzen, Cameron J.

Paper 1: Beyond Educational Attainment: Exploring Education Policy Predictors of Youth Voter Turnout. Individuals with higher levels of educational attainment are more likely to vote. Though this associational relationship is one of the most cited in political science, research has only recently confirmed that educational attainment increases voting. However, we still know little about which specific aspects of education matter—beyond years of formal schooling or degrees held. To advance beyond these measures, this paper explores whether state education policy variation in approaches to civics education, academic achievement, social and emotional learning, educational differentiation, and teacher resources can shed light on youth voter turnout (18 to 25 years-olds). Using state-level youth turnout data for national elections between 2010 and 2022 matched with a series of lagged education policy measures, this paper employs a series of two-way fixed effect regression models to explore whether state education policy can shape youth civic engagement. Results show that educational attainment predicts turnout at the state level, though not for young voters. Further, while most lagged measures for civic education, academic achievement, and educational differentiation exhibit no relationship with youth turnout, states with higher policy measures for social and emotional learning and education funding are consistently positively associated with higher turnout rates for youth. These findings illuminate the important dynamics of education that may shape voter turnout.

Paper 2: Explaining the Gender Gap in Voting: Civic Returns to Education.In recent decades in the United States, women have outpaced men both at the ballot box and in educational attainment. Since education is closely tied to political participation, this paper considers these two trends in tandem and assesses how much of the gender gap in voting is attributable to educational differences, differential returns to education, or other, non-education related elements. Using comprehensive educational data from Massachusetts students matched with voter records, this paper employs a Blinder-Oaxaca-Kitagawa decomposition to understand how educational attainment and other educational experiences contribute to gender voting differentials. In the sample, women outvote men by 3.85 percentage points in the first possible presidential election that young people can vote in after allowing time to complete college. Results demonstrate that two-thirds of this gap in voting is due to differences in educational attainment by gender, with only some of the remaining third of variation explained by either gendered differences in educational experiences or gendered returns to these educational characteristics. These findings broadly suggest that the gender gap in voting can be explained by a rise in women’s education and that if men reached the educational levels of women, they would vote at similar rates.

Paper 3: Navigating the Administrative Burdens of Voting: The Role of Education. The multitude of state election laws enacted in recent years implies a widespread acknowledgement that the “direct costs” of voting matter. Recent studies have affirmed that the costs of voting, such as those imposed by changes in election laws requiring voter identification, can reduce turnout particularly among certain groups. Other work has demonstrated laws that reduce the costs of voting do not always increase turnout. Amidst these conflicting findings, I argue that the impact of changes in the cost of voting are best understood in aggregate, as the process of voting in each state is governed by a web of overlapping laws and requirements. I further argue that increases to the direct costs of voting disproportionately impact areas with less-educated populations. Using two-way fixed effects models for county-level voter turnout in the six presidential elections between 2000 and 2020, I estimate that a standard deviation increase in the aggregate “costs of voting” decrease turnout in U.S. counties by roughly 1.3 percentage points. Additionally, interacting county education rates with changes in costs of voting show that the impacts are concentrated among counties with lower education levels—a one standard deviation increase in the cost of voting only decreases turnout by 0.86 percentage points in highly-educated counties of each state. Echoing work that shows such administrative burdens disproportionately affect less educated individuals, these findings offer suggestive evidence that increases in the costs of voting push less educated individuals out of the electorate.