Young adults have always been a prominent cog in American democracy. From the signing of the Declaration of Independence in 1776 to the recent student-led protests across the nation in most recent times, youth have been an outspoken voice for change.
Since ratifying the 26th Amendment in 1971, which decreased the voting age from 21 to 18, young voters have been turning out to the polls less and less, only to spike to record-setting levels in the 2020 presidential election. Why is this?
This question presented the opportunity to jump down a rabbit hole of questions that branched off each other. However, rather than focusing on influences that students had a choice with, such as attending protests and using social media, I wanted to view the effects on student engagement under the influence of things out of their control such as education policies.
According to CIRCLE, it was estimated that around 50 percent of voters in the United States, ages 18 to 29, participated in the 2020 election process. Despite the attractive increase in youth participation in elections, there has been a growing concern among policymakers about the lack of civic courses provided in schools across the nation. This lack of civic knowledge is broadcasted through local outbursts at State Canvasser meetings meant to establish election results to national security crises such as the January 6th insurrection.
Similar studies regarding youth participation in civics have been conducted.
Ashley Jeffrey and Scott Sargrad from the Center for American Progress’ study measures the requirements of civic education by state as of 2019 and assesses how the data impacts civic engagement among young adults. In their study, they determine that civic courses not only help citizens engage in civics outside of voting (such as protesting) but how civic courses also help citizens to be able to interpret political media and verify sources.
In his study of the effects of political partisanship on youth voter turnout in 2020, Vinay Khosla attempts to look at sociological factors that may deter youth voters from participating in elections. Of his 15 interviewees who fell between the ages of 18 and 19, he finds a positive correlation between high levels of polarization and high voter turnout.
If states require civic education courses for students to graduate high school, then citizens will be more likely to engage in civics post-graduation into early adulthood.
This study's dependent variable was civic engagement in early adulthood. I observed estimated youth voter turnout, ages 18 to 29, in the most recent elections, the 2016 National election, the 2018 Midterms, and the 2020 National election, across individual states within the United States. For this variable, I collected the data from CIRCLE for each of the election years.
The independent variable was civic education requirements. This was measured through levels of civic education or participation required of high school students to graduate in individual states. When observing the data collected by the National Center for Education Statistics, any state that had a required civics course was coded as 1, and those that did not require a civic course for graduation were coded as 0. Though it should be noted that a state did not have to require a strict government course to qualify as requiring a civics course – many were coded as 1 if within their History course it was required that students are taught about state or national government and its workings. Additionally, though available, I did not control for the number of credits required by each state nor other factors available within the variable.
Once I had observed how my independent and dependent variables worked together, as I will discuss later, I incorporated my control variables. My first control variable was ‘Facilitative Laws’ in each state. The facilitative laws data from the 2016 and 2018 election years were pulled from CIRCLE. CIRCLE gave a numerical score between 1 and 9 that measured how strongly a state had implemented laws that facilitate registration and voting, such as pre-registration and automatic registration. There was a total of 9 points total, in which states would receive one point if the law or policy had been enacted, and an additional point if that law or policy would be in effect by November of that election year.
My final control variable was Median Household Income across states in election years. This data from 2016 and 2018 was pulled from the United States Census Bureau.
The most important step of my study was to first compare how youth voter turnout fluctuated between 2016 to 2018, 2018 to 2020, and 2016 to 2020 without controlling for any other variables. This would set the precedent for the rest of my study as it would allow me to observe how much of a difference in voter turnout there would be between national election years versus midterm election years.
I found that there was an observable decrease in youth turnout between the 2016 presidential elections and the 2018 gubernatorial and legislative elections by 12%. This was not unexpected, as average voter turnout regardless of age tends to drop in individual state elections. However, when I compared youth voter turnout in the 2016 election to youth voter turnout in the 2020 election, there was an increase of turnout by 9.5%. This aligned with what the various other points of research that have been referenced claimed to observe.
After identifying how the dependent variable worked without influence, I added the independent variable of whether a state required some sort of civic education program in order for a student to graduate high school. In my observations, I found that both states that required a civics course to be taken and those that did not require a civics course to be taken showed dispersion. However, states that required no civics course to be taken showed more dispersion between years of data collection.
For example, between 2016 and 2018 there is a 2.72-point increase in standard deviation within states that do not require a civics course and a 1.32-point decrease in standard deviation within states that do require a civics course. Between 2018 and 2020 there is a 1.68-point increase in standard deviation within states that do not require a civics course and a 1.42-point increase in standard deviation within states that do require a civics course. I then compared the 2016 results to the 2020 results (as these were both national election years) and found that there is a 4.40-point increase in standard deviation within states that do not require a civics course and a 0.10-point increase in standard deviation within states that do require a civics course.
This finding follows with my observations of my dependent variable alone– where there will be a considerable decrease in 2018 due to lower levels of voter turnout.
I then began to incorporate my control variable of States with Facilitative Laws.
Unfortunately, due to the lack of data available 2020 results had to be excluded from this study. When applying the control variable for each respective election year, I found that from the 2016 to 2018 elections, the presence of facilitative laws in individual states presented an increase in the p-value (0.070 in 2016 to 0.080 in 2018). However, I had to keep in mind that there would already be a present decrease as preluded in observations of the dependent variable on its own in the decline of voter turnout. Because both the 2016 and 2018 p-values fail to fall below the threshold of 0.05, I did have to reject the hypothesis. Despite having to reject the null hypothesis, that does not conclude that there is absolutely no correlation between these two variables as each is respectively close to the threshold.
Finally, I made a scatterplot comparing Youth Voter Turnout and Median Household Incomes for both 2016 and 2018, once again excluding observations from the 2020 election year due to a lack of data.
Both the results in 2016 and 2018 show a positive correlation between voter turnout and median household income, implying that as a state’s median household income increases, so does youth voter turnout within that state. Additionally, both p-values (0.010 for 2016 and 0.004 for 2018) fall below the 0.05 threshold, therefore rejecting the null hypothesis and confirming there is a statistically significant relationship between these two variables.
This study was interesting to me as it both settled my theories about civic education and its impact on youth voter engagement, but also showed several other options that could further explain my hypothesis.
Based on my findings in this project alone, I conclude that civic education does not have a significant effect on youth voter turnout.
However, this only spurs the ideas of ‘what if?’ What would happen if I controlled for how much a state spent on student civics courses? What if I controlled for the number of credit hours required? More election years? Race? Gender? Religion? These are all branches of this that I was unable to explore in-depth, but I believe are valid in further determining an extensive conclusion for this hypothesis.