Vote Early, Vote Often

The role of schools in creating civic norms

It has been an almost uncontested proposition since the founding of the republic that America’s schools have a duty to prepare young people for active citizenship. As Thomas Jefferson put it, in arguing for a national system of schools, the idea was to have a common curriculum where “a foundation [would be] laid for a government truly republican.”

Today 40 state constitutions explicitly refer to the need for an informed electorate; 13 of them, according to a 2003 Carnegie Corporation report, “state that the central purpose of their educational system is to promote good citizenship, democracy, and free government.”

The problem is, despite Jefferson and the good intentions of state constitutions, we still don’t know exactly how schools are supposed to nurture this enlightened civic engagement. Most young people simply don’t vote, violating the first commandment of civic duty.

In the 1972 presidential election, when the ink was barely dry on the Twenty-Sixth Amendment (lowering the voting age to 18), only 52 percent of those aged 18-24 showed up at the polls (see Figure 1). Twenty-eight years later, in the closest presidential race in modern history, youth turnout, having fallen steadily over the years, reached a new low of 36 percent. Despite unprecedented efforts to mobilize younger voters for the 2004 presidential election, the turnout rate in the 18-24 age range was still only 45 percent-higher than in 2000, but nonetheless just at the average through the 1970s and 1980s. In future years, it is likely that voter turnout overall will continue to decline given the current trajectory of civic engagement among the youngest members of the electorate.

If America’s schools have a mandate to prepare young people for a lifetime of active citizenship, the evidence seems to suggest that schools are failing. But why? What can be done?

Click for enlargement.
Click for enlargement.

What Brings People to the Polling Booth

To understand why some people vote while others do not, one must concede the obvious fact that the outcome of an election seldom turns on the vote of any one person. Influencing the outcome of an election is thus not by itself a good reason to take the time to vote. Though many nonvoters say they do not go to the polls because their vote does not count, research shows that this is an excuse, not a characteristic that distinguishes them from voters.

Those who vote, in fact, do so because of a sense of civic responsibility, not because they believe their votes will change an outcome. In their 1995 study of civic involvement, political scientists Sidney Verba, Kay Schlozman, and Henry Brady asked politically active Americans why they engage in an array of civic and political activities. An overwhelming majority reported doing so because they feel it is their duty. This is only the most recent and comprehensive of studies that, since the 1950s, have found a clear connection between a sense of civic obligation and political participation.

This means that voting has roots in communal life. It is an individual action taken for reasons that go beyond the immediate interests of the individual voter. Communally, voting is an indicator of a government’s legitimacy. What, after all, distinguishes a democracy from all other forms of governance if not the vote? As voter turnout climbs, the legitimacy of the republican enterprise is enhanced. Indeed, many commentators interpreted the higher-than-expected turnout for the January 2005 Iraqi vote as an indication of the new government’s legitimacy.

Apparently, then, those who vote have an intuitive sense of the legitimizing power of the vote. Do they learn it at home? In their community? Or perhaps in school? The fact is, until now, it wasn’t clear-not even for Americans.

Schools and Civic Obligation

Researchers looking for links between school and voting habits in the United States have typically focused on the impact of the school’s civics curriculum. Despite assiduous efforts, however, they have usually found no discernible relationship. In 1998, Richard Niemi and Jane Junn, in summarizing the research, concluded that the conventional wisdom among political scientists is that civics classes have no effect on young people’s political engagement. Some scholars, a group that includes Niemi and Junn, have tried to challenge that conclusion, but they can do no better than show that children who take civic classes do marginally better on civics tests while they are actually enrolled in civics classes than students who are not enrolled in the same classes. Not a particularly remarkable finding, and not one that sheds much light on the influence of schools on civic engagement.

A much more promising line of inquiry is to ask how experience in a school community inculcates an appreciation for the value of civic engagement. And that branch of inquiry has borne some fruit.

If people are voting because of a sense of duty, then it is reasonable to wonder where these civic norms come from and how they are communicated to individuals. It makes sense to inquire about the strength of norms within the community and the effect of those collective norms on an individual. As Stephen Knack, now a senior economist at the World Bank, explained in a 1992 essay, “Civic Norms, Social Sanctions, and Voter Turnout,” even people with a weak personal sense of obligation have a greater likelihood of voting in a place that is populated with duty-bound compatriots. They do this, as Knack points out, because “someone with a low sense of civic obligation may nonetheless vote to avoid displeasing a friend or relative with a stronger sense of duty.” What matters, then, is not just whether a given individual has a strong sense of duty, but also whether that individual is surrounded by others with that commitment.

Knack’s analysis owes much to the work of the celebrated sociologist James Coleman, who was among the first to use the concept of social capital in a systematic way. The norms within a community that facilitate cooperation, Coleman showed, pay dividends for the community as a whole. In a 1987 essay, “Norms as Social Capital,” Coleman noted that people act in accordance with norms to avoid social sanctions-a disapproving look, a raised eyebrow, the whispered label of “shirker”-as well as to earn approval. Such social rewards and punishments, he suggests, are internalized by the individual and become part of that person’s own ethic, which then redounds to the benefit of others.

Since adolescence is a particularly important period of life for the adoption of norms, schools can be expected to be an especially important incubator of norms of civic participation. This is why Coleman centered his work on social capital in schools. And if he is correct, we should find a link between strong civic norms at school and engagement in the political process, in both the present and the future. In short, the civic norms at your high school should affect whether you vote.


Although well-known for many years, these ideas have not been put to a careful empirical test, in part because few surveys of students contain detailed contextual information about their school community and track the students into adulthood. Fortunately, the University of Michigan’s pioneering Youth Studies Series (YSS) does both. The YSS began in 1965, when a representative sample of the nation’s high-school seniors and their parents were interviewed. These students were part of a panel, which means that they were surveyed again in later years-1973 and 1982. (Because the first wave of data was collected in 1965, just as the Voting Rights Act was beginning to be implemented and while many African-Americans were still deprived of the vote, I limit my analysis to whites only.)

Information was also gathered in 1965 from roughly 125 students from each sample member’s high-school class. These students answered a similar, but shorter, version of the questionnaire given to students who were in the panel. For every question asked of these additional students, average measures for their high schools can be calculated independent of the panel member’s own qualities. For example, the answers will show whether an individual student intends to attend college, and the aggregation of the entire school sample can give an estimate of the percentage of students within that school who plan on going to college.

My analysis of these data proceeds in three stages. First, I examine whether the civic norms in a high school in 1965 affected whether adolescents anticipated being politically engaged as adults. Second, I test whether the civic norms in the schools individuals attended in 1965 affected their likelihood of voting in the 1980 presidential election. In both cases, my central hypothesis is that the stronger the school’s civic climate, as measured by the strength of a school’s communal belief in voting, the greater the degree of electoral engagement for any of the school’s individual students, regardless of their own personal sense of civic duty. Having confirmed the independent influence of civic norms within a school, I turn finally to exploring conditions that facilitate their development.

Each stage of the analysis hinges on the availability of a reliable indicator of civic norms. The YSS makes it possible to gauge the civic norms within a high school with a question that asked students, “What three things about a person are most important in showing that he is a good citizen?” Students were asked to choose from a list of six options that included religious involvement, adherence to the law, a sense of privatism (or minding one’s own business), not considering oneself better than others, being proud of one’s country, and, most important for this study, voting in elections. The beauty of this measure is that it relies on the students’ initiative to identify voting as a component of good citizenship and avoids the problem of having everyone reflexively endorse voting as a normative expectation. As a result, the percentage of students in each high school who endorse voting as a component of good citizenship varies widely, ranging from 46 to 85 percent (with an average of 70 percent) across the 77 schools in the sample. I will refer to the percentage of students in each school who link voting with good citizenship as the school’s civic climate.

Anticipating Engagement

I begin by testing whether the strength of a school’s civic climate affected its students’ anticipation of being engaged in politics later in life. Specifically, students in the sample were asked, “Looking ahead to the time when you are on your own, what about actual participation in public affairs and politics? How active do you think you will be in these matters?” In response, they had three choices: not very active, somewhat active, very active. It is likely that, to an adolescent, this question is really asking, “Do you think people should be engaged?” Or, perhaps more accurately, “Do you think people like you should be engaged?”

In testing whether adolescents attending schools with a strong civic climate were more likely to envision themselves as active citizens, I control for differences in a wide variety of individual and contextual factors that could also affect anticipated engagement. These factors include, for the individual, the level of education the student planned to attain, the level of education the student’s parents attained, the length of time the student had lived in the community, and gender. I also control for differences in the average level of education attained by the parents of other students in the school, the average length of time that its students have lived in their community, and the political diversity within the school population (as measured by differences in self-reported major-party affiliation).

Perhaps most critically, I also take into account whether each individual endorsed voting as a sign of good citizenship. In other words, my analysis isolates the effect of being surrounded by others who see voting as a component of good citizenship regardless of the individual’s own expressed sense of civic duty. As expected, adolescents who reported that good citizens exercise their right to vote also reported that they anticipate being publicly engaged on reaching adulthood. All else being equal, endorsing voting as a mark of good citizenship boosted anticipated engagement by about one-eighth of a standard deviation-a slightly larger impact than resulted from an increase in parents’ education of one standard deviation.

Yet even when taking into account the individual’s own sense of responsibility, the school’s civic climate had its own impact on how students plan to live their political lives. Impressively, attending the school with the strongest civic climate (where 85 percent of students listed voting as a component of good citizenship) rather than the school with the weakest civic climate (where 46 percent chose voting) increased anticipated participation by a quarter of a standard deviation, or by about twice the effect of the individual student’s having listed voting as a component of good citizenship.

Meeting Expectations?

That more students in a school with a strong civic climate anticipate being engaged in politics is surely a good thing, but, of course, teenagers do not always live up to even their own expectations. Does the civic climate of one’s high school also affect the likelihood of turning out to vote in the years following high school?

To address this question, I use data from the second round of YSS follow-up interviews, conducted in 1982, which asked the study’s participants to report on whether they voted in the 1980 presidential election. This election occurred roughly 15 years following the panel members’ graduation from high school, when most were in their early 30s. The overall voter turnout that year, roughly 55 percent, was fairly typical for elections after 1968.

In examining the impact of the school’s civic climate, it is again important to take into account other factors, focusing this time on those most likely to influence the respondent’s decision to vote. At the individual level, this includes education, marital status, gender, and length of residence in the community-all factors, past research has shown, that have a bearing on turnout. And since the home is also a critical factor in a person’s civic development, I control for three measures of the engagement level of the student’s parents as of 1965, including whether the parents voted in the 1964 presidential election. Finally, I again take into account whether the individual respondent endorsed voting as an essential component of good citizenship in answering the question used to gauge the school’s civic climate in 1965.

I also control for several characteristics of the respondent’s high school, including the average level of education attained by the parents of other students in the school, average residential stability, average level of anticipated participation, and partisan diversity within the school environment. Of course, for the purpose of testing the impact of civic norms, the most significant school-level variable is its civic climate: the percentage of students who viewed voting as an obligation of citizenship.

As in the case of students’ anticipated participation, the civic climate of the school has a clear impact. People who in 1965 attended a school where students saw voting as a civic duty were far more likely to vote 15 years later. All else being equal, a typical individual who attended the school with the largest percentage of students identifying voting as a component of good citizenship was 14 percentage points more likely to vote than the person who attended schools with the lowest percentage of students identifying voting as a civic duty. By way of comparison, having a college degree increased the likelihood of voting by 18 percentage points. An individual who attended a school with a civic climate one standard deviation above the mean school in the sample was about 7 percentage points more likely to vote than a similar individual attending a school with a civic climate one standard deviation below the mean (see Figure 2). Meanwhile, the partisan diversity of an individual’s school had no impact whatsoever on later turnout-a finding that is important to keep in mind when interpreting the results presented later in the story.

After accounting for the school’s civic climate, the fact that an individual student identified voting as a civic duty in high school also has no effect on voting later in life. However, had I not measured the effect of the school’s civic climate, I likely would have concluded that an individual’s attitudes toward voting as an adolescent were quite important. In doing so, I would have attributed to the individual what was in fact a consequence of the environment in which he or she was educated.

Click for enlargement.

Click for enlargement.

A Civic Norms Nursery

Finding that communal civic norms in high school lead to voter turnout years later leads naturally to the question of what conditions facilitate the incubation and nurturing of those collective norms within a school. Myriad possibilities present themselves, most of which are beyond the scope of these data to test. The findings of scholars who study social capital, however, suggest one possibility that can be tested: homogeneity-whether racial, religious, or economic-seems to foster social capital. Or, to put it more provocatively, diversity may diminish the social cohesiveness necessary to sustain strong civic norms.

With this in mind, I examine how diversity influences the civic climate of a high school, focusing on three distinct dimensions of that diversity: racial, religious, and partisan. Although most research to date on the effects of diversity has focused on demographic characteristics like race and religion, there is good reason to think that political diversity may be even more important-especially with regard to norms about voting. Recall that the theoretical link between cohesiveness and social capital is that of widely shared values. Those shared values constitute a cultural outlook that research shows is often expressed through political party affiliation. Partisanship, then, provides a particularly good proxy for shared values, even though high-school students are unlikely to identify each other by partisan labels.

As hypothesized, I find that the more politically homogeneous the high school, the stronger the norm linking voting with good citizenship (see Figure 3). This is true even after taking into account two other factors that might be expected to affect a school’s civic climate; namely, the mean level of education attained by students’ parents and the average length of time that students have lived in their community. The impact of political homogeneity on civic climate is about the same as that of the average parents’ education, which is noteworthy given that education and the status it confers has long been recognized as a major facilitator of social norms generally and of civic engagement specifically. Only political diversity adversely affects schools’ civic climate; neither religious nor racial diversity has effect.


The bottom line of this analysis is that the civic climate in high school has a great impact on voter turnout at least 15 years following graduation. What matters is that an adolescent’s community, defined in this case as the high school, is populated with a high percentage of peers who express their belief that voting is an indicator of good citizenship. In fact, after accounting for the civic climate of an individual’s high school, an individual’s own belief that voting is a civic duty does not have an impact on voting as an adult. Individuals do not act, nor are they acted on, in isolation. Rather, norms are inculcated within communities, such as the family, the neighborhood, and the school. And those norms reach into the future.

I also found that cohesive schools, including those with a homogeneous political composition, foster civic norms. This is clearly an explosive finding: diversity, at least along a political dimension, dampens civic norms. A naive reading of this result might suggest that the cure for America’s civic ills lies in crafting clusters of political homogeneity, with conservatives shunted off to their schools and liberals to theirs. Such an inference would be dead wrong. Recall that political cohesiveness had no discernible effect on adolescents’ later voting patterns after the civic climate within the school was taken into account. And surely there are many factors apart from political diversity that can influence civic norms.

A more sensible conclusion to draw is that any school-based reform aiming to enhance voter turnout among the rising generation should focus on ways to foster a strong civic climate. The importance of political cohesiveness provides a clue to how to do this. Political cohesiveness is merely an indicator of values held in common within a community. A promising course of action to strengthen a school’s civic climate, therefore, is to identify ways to foster a sense of commonality among a school’s students.

For this to happen, legislators, educators, parents, and the public must recognize that the civic dimension of our education system deserves more than lip service, but should be subject to the same scrutiny as other education outcomes. Schools will take civic education seriously only when policymakers and parents begin to scrutinize the civic experiences provided by their schools as closely as they currently monitor academic standards. I echo political theorist Stephen Macedo’s lament: “Given the centrality of civic purposes to public schools it is ironic that studies of -effective schools’ pay so little attention to civic ends.” We need to learn what works and then make it possible for our schools to do it.

But is it realistic to expect today’s schools, beleaguered as many of them are, to add the promotion of civic duty to their long list of responsibilities? I believe that it is, and that an unlikely model can be found in the way American schools have in recent decades come to embrace tolerance for diversity as a preeminent value. This respect for differences has not come about through a particular class that students take, nor is it restricted to a select few who participate in a specific program. It has become part of the modern public school’s culture, quietly and consistently reinforced by teachers and administrators. It has, in other words, become a norm.

Tolerance for diversity is a noble objective, and America’s educators should be applauded for having successfully integrated it into the education of today’s students. However, while tolerance is necessary for a vibrant democratic culture, it is not sufficient. Turnout among the youngest voters has declined, even as they have grown more tolerant; plainly, a sense of civic duty is also needed. So let’s consider the way tolerance has become a norm as a template for a renewed focus on encouraging a sense of civic duty among today’s adolescents. Our schools need to reemphasize America’s shared civic culture, not to replace but to complement encouraging tolerance. Let educators do for unum what they have done for pluribus.

-David E. Campbell is an assistant professor of political science at the University of Notre Dame. Article funded by a grant from the John M. Templeton Foundation.

Last Updated


Notify Me When Education Next

Posts a Big Story

Business + Editorial Office

Program on Education Policy and Governance
Harvard Kennedy School
79 JFK Street, Cambridge, MA 02138
Phone (617) 496-5488
Fax (617) 496-4428

For subscription service to the printed journal
Phone (617) 496-5488

Copyright © 2024 President & Fellows of Harvard College