Introduction
Over the last few decades, many research studies have been conducted to determine whether higher education impacts voter turnout in the United States. In general, it is widely believed that Americans with higher education vote at higher rates, which suggests a link between years of schooling and voting. But does achieving higher levels of education cause citizens to show up and vote on election day? Or does education and voting simply go hand-in-hand because some other variable contributes to both? The research on this topic is mixed, with some studies confirming evidence of a causal relationship, while others are not so certain. In this study, I will dive into exploring the effects of higher education, as it is increasingly becoming one of the central topics up for discussion when analyzing voter turnout. My study will investigate and ask the key question: What is the relationship between higher education and voter turnout? I will examine the levels of statewide higher education (bachelor’s degree or higher) as my independent variable and test whether or not it has an observable impact on the rate of voter turnout per the population of each state, my dependent variable. The hypothesis that higher education levels will not influence electoral participation, nor affect voter turnout, in the United States. I will be honest and upfront in my results, as a better understanding of how measures work to impact or not impact voter turnout will help to better inform future legislators. If higher education has no effect, then other factors will need to be considered; however, if higher education has an observable effect, then this knowledge will be much appreciated going forward.
Literature Review
Early research has shown that high education levels among individuals in the United States will significantly increase the number of people turning out to vote in local, state, and national elections (Powell, 1986). Evidence has revealed that having a solid educational background will instill a sense of civic duty, political efficacy, and political sophistication and awareness (Jackson, 1995). Other past evidence has shown that there have been considerable declines in voter turnout rates when education levels have also seen declines over the years (Boyd, 1981). It would stand to reason that a fall in education levels could possibly result in a fall in electoral participation, as less informed groups of people will emerge holding little knowledge or expertise about how the electoral process works and their potential roles across the political spectrum. If individuals continue to receive high levels of education, there is less expectation for low voter turnout rates across the United States. In 2014, only 47% of 18–24-year-olds were registered to vote, the lowest in 40 years (Berinsky, 2011). But, despite the fact that voters under the age of 30 participate at much lower rates than the general population, researchers in the past have estimated that there is a significant positive relationship between taking a civics course and extracurricular participation in high school and the likelihood of voting as a young adult, even after adjusting for other causes of civics education and voter turnout (Berinsky, 2011). Similarly, having a college-level education can indirectly increase voting since degree attainment and rising socio-economic status will promote political participation, as observed during the 2004 presidential election when 54% of the nation’s youngest cohort of voters(18–25-year-olds) cast their votes (Fernandez, 2021). I will further explore the past literature on this topic, including how the authors have gone about measuring the effects of political education, the main takeaways from their studies, and the lens through which they view value and progress in determining the true results of higher education and its impact on electoral participation and voter turnout in the United States.
Secondary Educations
A key component of having a higher education is the probability that individuals will be better informed about the electoral process and how their citizen participation in the government will play a critical role across society. It is not surprising, then, that there are many studies centered on discussing the feasible political benefits of higher education. Does having a solid knowledge of politics enforce individuals’desire to participate and vote in elections once they come of age? The early research behind this question seems to be widely unanimous on the point that higher secondary education levels do result in an increase in political involvement. In terms of having an impact on political performance, the general public is in agreement that individuals with solid secondary educational backgrounds will participate to a larger extent in political activities than individuals with lower secondary educational backgrounds (Berinsky, 2011). In this study, a propensity-score matching process was used to take into account pre-adult experiences and influences in place during students’ senior year of high school. The results revealed that secondary education is the precursor for pre-adult experiences, showing a direct causal link to eventual political participation and voter turnout (Berinsky, 2011).
College Educations
Turning away from research focused on the effects of secondary education to more recent research that examines the effects of college-level education, I was faced with more in-depth studies that argue higher education can influence political behavior and encourage voter turnout across the United States. Out of the research that focuses specifically on the effects of college-level education, some researchers have argued that the apparent link between education and voter turnout is spurious, while other studies have found that colleges seek not only to increase student voter turnout but also to prepare students to become informed voters (Fernandez, 2021). In this study, matching survey procedures were used to account for pre-college civic engagement and estimated treatment effects of taking at least one community college course. The findings demonstrated that taking political science influences college student voter registration, voter turnout, and the ability to correctly answer questions about the roles of U.S. political institutions (Fernandez, 2021). Another study conducted in 2009 drew similar conclusions, but approached the question by analyzing two randomized experiments and one quasi-experiment in which educational attainment was altered exogenously (Sondheimer, 2009). Students in these experiments were analyzed over the long-term, eventually examining their voting rates as adults and proving that college graduation powerfully affected voter turnout, indicating that the relationship between education and turnout was indeed causal (Sondheimer, 2009). Some studies fall in the middle of these debates, seeing both the benefits and detriments of higher education on voter turnout. A 2014 campus field experience on a small, private university in Florida found that although face-to-face political mobilization efforts positively boosted voter turnout, the university had overall made little progress in nurturing political engagement (Hill, 2014). The authors found a non-significant difference in turnout between the students contacted and those not contacted, suggesting that while generally difficult to mobilize U.S. citizens, it may be especially challenging to mobilize U.S. college students while they are still at school (Hill, 2014).
Political Behaviors
Additional studies help to back up the positive beliefs surrounding the relationship between political education and political behavior. One study from 1998 utilized survey data from 1996 to examine the extent of political knowledge in Australia and analyze its consequences for political literacy, competence, and participation (McCallister, 1998). The results revealed that the median citizen could answer 2 out of 7 objective questions correctly, with women, adolescents, and those with less education being more likely to provide incorrect answers (McCallister, 1998). The association between knowledge, attitudes, and behaviors showed that factual knowledge increases political literacy, making the political knowledge gained through education in schools more effective in generating positive views of democratic institutions and thus shaping the political behaviors individuals already possess (McCallister, 1998).
In summation, different research studies have been conducted over the years on the topic of higher education and its impact on political participation and voting in the United States. It has been generally agreed upon in these studies that higher education is beneficial to foster political understanding and involvement, but whether or not these perceptions are true is largely debatable, and the rationale for why this is the case is still open to investigation. The current research does not deliver a broad enough consensus of participants, which is a limitation of the examinations. All the studies consisted of narrow and parallel experiments involving individuals as their units of analysis, typically groups of students or single universities. In this study, I will be broadening the scope of research, using states as a unit of analysis to observe the percentage of states with a bachelor’s degree or higher, and comparing the data to the percentage of voter turnout in the United States. I will also be looking at the rate of poverty-level living in each state and the percentage of Republican leanings to draw further conclusions about my data. This approach will present a larger and broader representation of the relationship between higher education and voter turnout. By examining statewide trends, this topic will be investigated in a new, uncharted way that has not been explored in other existing research methods.
Methodology
To begin, the operational definitions will be identified for all the variables used to test my hypothesis. The independent variable is the levels of statewide higher education. It is a ratio variable, measured as the percentage of bachelor’s degree or higher education levels in the United States. The data regarding whether a state has higher education or not was collected by the World Population Review (WPR). The WPR obtains its data by gathering population statistics from the United Nations Population Division’s “World Population Prospects,” which is a comprehensive collection of population estimates centered on national census data, vital registration systems, and sample surveys conducted around the world. This is a good source because it provides broad, up-to-date global population statistics and yearly education trends in an accessible format, drawing from reliable government resources.
The dependent variable is voter turnout per population by state. It is a ratio variable and is per population, in order to organize all the states in the same way. When calculating this ratio, the denominator (the base of the calculation) is always based on the total population size within a specific state, allowing for meaningful comparisons between different populations, even if their absolute numbers differ. This data was also collected by the World Population Review (WPR). The WPR collected the data from official updated reports on recent trends in voter turnout across the United States.
The first control variable is the poverty rate. Poverty rate is a ratio variable and is measured as the percentage of poverty-level living per the population of each state (numbers vary depending on family size). This data was collected by the World Population Review (WPR). The WPR collected this information from government-reported income and tax data.
The second control variable is Republican Party preference, a ratio variable associated with this study. It is being measured as the percentage of Republican leanings per state. This data was also collected by the World Population Review (WPR), which provides clear and updated statistics on state-wide party affiliations.
For each variable, the unit of analysis is thestates. My study is examining the attributes of states in each of the four variables. There is no sampling to be drawn since this investigation would be most effective by looking at the entire population of states. I will conduct the appropriate tests for ratio variables to examine the levels of statewide higher education (IV) and voter turnout (DV), to ideally compare the results of the two variables and determine meaningful outcomes. As a result, this requires a population.
All the measures can be deemed quite valid and reliable for the most part. Whether graphically showcasing the percentage of statewide higher education or simply displaying and interpreting population statistics, the data is coming from a direct and easy-to-locate source that can be verified for legitimacy to generate consistent results in repeated tests. However, it is important to keep in mind one potential concern about this data set. It is the fact that the World Population Review (WPR), from which the data for my independent, dependent, and control variables originated, was independently attained. The WPR is one of the most trusted, reliable, and up-to-date aggregators of statewide statistical data, but it has no formal oversight and is still a private entity without political affiliations or a specific agenda. As a result, that background is certainly something to consider.
The descriptive statistics of my variables will be tested. These calculations will give the mean, median, mode, range, interquartile range, variance, and standard deviation of my dependent and independent variables, determining the amount of voter turnout per population by state, as well as revealing the proportion of states with higher education. These figures will lead to the linear regression analysis test, which will help to identify if a causal relationship between the four variables exists by giving a p-value and significance level. If the p-value determines significance at the 0.01 level, a potential relationship between the variables can be suggested. Secondly, a correlation coefficient test will initially be performed between just my independent and dependent variation will be observed. If the results gathered are close to 1, there is a strong, positive relationship, and if the values found are close to -1, there is a strong, negative correlation between the independent and dependent variables. The outcomes will vary to some extent, but will show if there is any correlation between the two core variables. Next, I will generate a similar regression test in the form of a table that analyzes my two control variables in relation to the independent variable. Finally, a correlation coefficient test representing all four of my ratio variables will be conducted. If the results gathered are close to 1, there is a strong, positive relationship, and if the values found are close to -1, there is a strong, negative correlation between the control variables and the independent and dependent variables. Again, the outcomes will vary to some extent, but will show which, if any, controls are correlated.
Every research study contains limitations. Two key limits exist in this study. The first concerns the data, and the second involves the research design. First, as previously mentioned, the data for my independent, dependent, and control variables were collected from a private entity that is widely considered one of the most reliable and updated statistical sources, but there is still room for improvement. The data could be a little vague, as the WPR has no formal oversight or specific agenda. Second, the research design is limited by the fact that all the variables are ratios. Overall, using all ratio variables is considered quite advantageous due to their high level of measurement. However, there is a small possibility that using only ratio variables could oversimplify complex relationships and fail to capture the nuances of human behavior, potentially leading to inaccurate parameter estimates and reduced statistical power. Perhaps, as more research is carried out, the tests can become more varied and subsequently more valid. But for now, I must work with the measurements I have to draw conclusions and garner results.
Results and Discussion
All visualizations showing the relationships between my four variables can be reviewed in the Appendix. Figure 1 (see Appendix A) is a boxplot that illustrates the independent and dependent variables’ central tendency, spread, and potential outliers. The boxplot displays the median, quartiles, and extreme values (minimum and maximum) to provide a concise overview of the data distribution. Figure 2 (see Appendix B) demonstrates the findings for the central tendency measurements. Figure 3 (see Appendix C) shows the findings for the measures of variation (min, max, range, Q1, Q3, interquartile range, variance, and standard deviation). These descriptive statistics can reveal a great deal about my data. There was found to be low variance (0.003) between higher education (bachelor’s degree or higher) and voter turnout, which indicates a high degree of consistency and close uniformity in my data. For example, the average voter turnout rate was 67.95%, and the median was 68.13%, with Massachusetts having the highest level of education and one of the highest rates for voter turnout at 72.11%. In contrast, the range for higher education was 23.30%, with West Virginia having the lowest level of higher education and also one of the lowest rates of voter turnout at 57.58%. Hence, after observing this initial data, I can predict that there is a clear correlation between having a higher education and high voter turnout rates.
More testing is needed to help confirm this assumption, which leads to my first test – the linear regression analysis. The scatterplot for this test can be viewed in Figure 4 (see Appendix D), and the numerical results can be observed in Figure 5 (see Appendix E). The trendline on the scatterplot indicates that there is a positive slope. For this data, the positive slope means that as the X variable increases, the Y variable tends to increase as well. For every one percent increase in levels of statewide higher education, voter turnout per population by state increases by 0.7264 percent. The multiple R on the numerical results confirms this, as there is a 66% strong, positive correlation between the variables. This association is robust, as the R-squared value is 0.4359. This tells me that Levels of Statewide Higher Education (X) explains 43.5927% of the variation in Voter Turnout Per Population By State (Y). Finally, a test for significance is run at the 0.01 level of probability. The p-value result of 1.82298E-7 shows that the regression model is statistically significant at the 1% level of significance. This evidence suggests that a state’s level of higher education does indeed impact the rate of voter turnout per the population of each state, as a meaningful relationship is present between the variables. It’s still important to note the correlation versus causation rule. For this data, the statistical significance indicates a non-zero correlation, but not necessarily a causal relationship.
A second test was conducted to further analyze the relationship between the independent and dependent variables. A correlation coefficient test was run to interpret the levels of statewide higher education and voter turnout per population by state. The results of this test can be observed in Figure 6 (see Appendix F). In connection with the dependent variable, I found that levels of statewide higher education and voter turnout per population by state are positively correlated at 0.66. This test reveals a strong, positive correlation between higher education and voter turnout. For this data, I can observe that when higher education increases, the rate of voter turnout increases as well. This interpretation has already been graphed out to visually interpret the correlation in the first test, which can be viewed in Figure 4 (see Appendix D).
Finally, I explored my two control variables. A regression analysis was conducted on the poverty rate and the Republican Party’s impact on voter turnout. The table for this test can be viewed in Figure 7 (Appendix G). The table displays a statistical regression analysis comparing two models that examine the factors influencing state voter turnout, my dependent variable. Model 1 only considers “Education” as a predictor, while Model 2 includes “Education,” “Poverty,” and “Party.” Education has a positive and highly significant effect on voter turnout in both models. Poverty has a significant negative effect on voter turnout in Model 2. Model 2 has a higher Adjusted R² value (0.475) than Model 1 (0.424), indicating that it explains more of the variance in voter turnout. Both models are statistically significant overall, as indicated by the F-statistics. Specifically, the table shows me how the coefficient for “Poverty” is -0.916, and it has two asterisks (**), with the note at the bottom of the table indicating that ** p < 0.05. This means that for each unit increase in the poverty variable, the state voter turnout is predicted to decrease by 0.916 percentage points, and this result is statistically significant at the 0.05 level. The results seem reasonable since higher poverty rates often correlate with lower voter turnout due to various factors, including resource constraints, reduced civic engagement opportunities, and feelings of political disenfranchisement among low-income populations. Furthermore, the table shows how the coefficient for “Party” (Republican) is -0.026, and it has no asterisks. This indicates that the effect of the party variable on state voter turnout is not statistically significant at the a = 0.1 level, as even a single asterisk for p < 0.1 is absent. Therefore, based on this model, there is no evidence of a relationship between the party variable and state voter turnout. The lack of statistical significance suggests that, within the context of this specific model and dataset, the party variable does not sufficiently explain variations in voter turnout. The actual impact might be negligible, or the variable itself might not be capturing a relevant aspect of party influence on turnout.
Additionally, a correlation coefficient test was conducted on all four variables to specifically interpret both poverty rate and Republican Party preference in correlation with the independent and dependent variables. Similar results in connection with the regression table were gathered. The results of this test can be observed in Figure 8 (see Appendix H). In connection to the dependent variable, I can see that poverty rate and voter turnout per population by state have a strong, negative correlation at -0.66. It makes sense for this to be true, as it is less likely to see substantial voter turnout in more poverty-stricken areas. I can also observe that voter turnout and Republican Party preference have a weak, negative correlation at -0.30. This also makes sense because not every voter will be supporting the Republican Party; some will also be voting Democrat, Independent, or have no party preference at all. The control variables’ regression analysis has been graphed out to visually interpret their negative correlation in Figures 9 and 10 (see Appendix I and Appendix J). I can confirm that these correlations are much weaker than the strong, positive correlation I observed between higher education (IV) and voter turnout (DV). As this study has proved my hypothesis to be wrong, further analysis on the implications of higher education (bachelor’s degree or higher) on statewide voter turnout could be very informative and instrumental to approaching the question of higher education’s effect on voting in the United States.
Conclusion
In conclusion, the goal of this study was to use new methods to answer the question: What is the relationship between higher education and voter turnout? By using states as a unit of analysis, I was able to examine much broader, nationwide research compared to previous studies that focused on just groups of individuals or single universities. However, I acknowledged that this study faced limitations in the fact that the World Population Review (WPR), from which the data for my four variables originated, was independently attained. This background raised the concern that my data could be slightly vague, as the WPR has no formal oversight or specific agenda. I hypothesized that higher education levels would not influence electoral participation, having no effect on voter turnout in the United States. This hypothesis was proven wrong, as the tests – linear regression analysis and correlation coefficient tests, as well as data calculations and visualizations – demonstrated consistent results showing a state’s level of higher education and how it impacts the rate of voter turnout per the population of each state. Furthermore, analyzing my control variables allowed me to see a strong, negative correlation between poverty rate and my dependent variable, and a weak, negative correlation between Republican Party preference and the dependent variable. As a final point, my study has provided clear evidence that higher education is a powerful, though complex, driver of voter participation, with profound implications for the future of democratic engagement. My research highlights higher education’s substantial civic benefits, suggesting that expanding access to college can serve as a crucial mechanism for reducing inequality in political participation and fostering a more representative electorate. As a final point, I hope that my study’s findings encourage election officials to take into account the positive impact higher education is currently having on voter turnout in the United States.
Works Cited
1. Berinsky, A., et al. (2011). Education and Political Participation: Exploring the Causal Link. Political Behavior, 33: 357–373. https://doi.org/10.1007/s11109-010-9134-9https://doi.org/10.1007/s11109-010-9134-9
2. Boyd, R. (1981). Decline of U.S. Voter Turnout: Structural Explanations. American Politics Quarterly, 9(2): 133–159. https://doi.org/10.1177/1532673X8100900201https://doi.org/10.1177/1532673X8100900201
3. Fernandez, F. (2021). Turnout for What? Do Colleges Prepare Informed Voters? Educational Researcher, 50(9): 677–678. https://doi.org/10.3102/0013189X211045982https://doi.org/10.3102/0013189X211045982
4. Hill, D., et al. (2014). Can Face-to-Face Mobilization Boost Student Voter Turnout? Results of a Campus Field Experiment. Journal of Higher Education Outreach and Development, 18(1): 61–88. https://openjournals.libs.uga.edu/jheoe/article/view/1093https://openjournals.libs.uga.edu/jheoe/article/view/1093
5. Jackson, R. A. (1995). Clarifying the Relationship Between Education and Turnout. American Politics Quarterly, 23(3): 279–299. https://doi.org/10.1177/1532673X9502300302https://doi.org/10.1177/1532673X9502300302
6. McCallister, I. (1998). Civic Education and Political Knowledge in Australia. Australian Journal of Political Science, 33(1): 7–23. https://doi.org/10.1080/10361149850697https://doi.org/10.1080/10361149850697
7. Powell GB. (1986). American Voter Turnout in Comparative Perspective. American Political Science Review, 80(1):17–43. http://doi:10.2307/1957082http://doi10.2307/1957082
8. Sondheimer, R., et al. (2009). Using Experiments to Estimate the Effects of Education on Voter Turnout. American Journal of Political Science, 54(1): 174–189. https://doi.org/10.1111/j.1540-5907.2009.00425.xhttps://doi.org/10.1111/j.1540-5907.2009.00425.x
9. World Population Review. (2025). Educational Attainment By State. World Population Review. https://worldpopulationreview.com/state-rankings/educational-attainment-by-statehttps://worldpopulationreview.com/state-rankings/educational-attainment-by-state
10. World Population Review (2025). Voter Turnout By State. World Population Review. https://worldpopulationreview.com/state-rankings/voter-turnout-by-statehttps://worldpopulationreview.com/state-rankings/voter-turnout-by-state
11. World Population Review (2025). Poverty Rate By State. World Population Review. https://worldpopulationreview.com/state-rankings/poverty-rate-by-statehttps://worldpopulationreview.com/state-rankings/poverty-rate-by-state
12. World Population Review (2025). Political Parties By State. World Population Review. https://worldpopulationreview.com/state-rankings/political-parties-by-statehttps://worldpopulationreview.com/state-rankings/political-parties-by-state
Appendix
Boxplot.
Central Tendency Measurements (Numerical Results).
Voter Turnout Per Population By State (DV) |
|
Mean |
67.96% |
Median |
68.13% |
Mode |
#N/A |
Levels of Statewide Higher Education (IV) |
|
Mean |
34% |
Median |
33.65% |
Mode |
0.278 |
Measures of Variation (Numerical Results).
Voter Turnout Per Population By State (DV) |
|
Min |
55.0% |
Max |
80.0% |
Range |
25.0% |
Q1 |
0.645 |
Q3 |
0.728 |
Interquartile Range |
-0.083 |
Variance |
0.00346742 |
Standard Deviation |
0.05888482 |
|
Levels of Statewide Higher Education (IV) |
|
Min |
23.30% |
Max |
46.60% |
Range |
23.30% |
Q1 |
0.309 |
Q3 |
0.37125 |
Interquartile Range |
-0.06225 |
Variance |
0.003 |
Standard Deviation |
0.054 |
Scatterplot (Regression Analysis, IV vs. DV).
Regression Statistics |
|
|---|---|
Multiple R |
0.660247725 |
R Square |
0.435927059 |
Adjusted R Square |
0.424175539 |
Standard Error |
0.045137276 |
Observations |
50 |
Linear Regression, Numerical Results.
Coefficients |
Standard Error |
t Stat |
P-value |
|
|---|---|---|---|---|
Intercept |
0.432623476 |
0.041056626 |
10.53724 |
4.44083E-14 |
Levels of Statewide Higher Education |
0.726396883 |
0.119265291 |
6.090597 |
1.82298E-07 |
Correlation Coefficient Test Between IV and DV.
Correlation Coefficient Results |
Levels of Statewide Higher Education (Ratio) |
Voter Turnout Per Population By State (Ratio) |
|---|---|---|
Levels of Statewide Higher Education |
1 |
|
Voter Turnout Per Population By State |
0.660247725 |
Regression Analysis Table.
Model (1) |
Model (2) |
|
|---|---|---|
Education |
0.726*** |
0.375* |
(0.119) |
(0.214) |
|
Poverty |
−0.916** |
|
(0.372) |
||
Party |
−0.026 |
|
(0.106) |
||
Constant |
43.262*** |
67.757*** |
(4.106) |
(14.092) |
|
Observations |
50 |
50 |
R2 |
0.436 |
0.507 |
Adjusted R2 |
0.424 |
0.475 |
Residual Std. Error |
4.514 (df = 48) |
4.311 (df = 46) |
F Statistic |
37.095*** (df = 1; 48) |
15.761*** (df = 3; 46) |
Complete Correlation Coefficient Test
Correlation Coefficient Results |
Levels of Stotewide Higher Educotion (Ratio) |
Voter Turnout Per Population By State (Ratio) |
Poverty Rote (Rutiu) |
Reoubiitan Porty Preference (Rotia) |
|---|---|---|---|---|
Levels of Statewide Higher Education |
1 |
|||
Voter Tumnout Per Population By State |
0.660247725 |
1 |
||
Poverty Rate |
–0.737764408 |
–0.666074505 |
1 |
|
Republican Derty Preference |
–0.552095679 |
–0.300768427 |
0.194565681 |
1 |
Scatterplot (Regression Analysis, CV vs. DV).
Scatterplot (Regression Analysis, CV vs. DV).