Chapter 3 • Turnout and Participation in Elections

IN EVERY election cycle, stories in the media question who and how many will vote. Are Trump voters more fired up than Biden voters, or is it the other way around? Will Hispanic voter turnout be higher than normal in reaction to Trump’s immigration policies? Will Black people vote at a higher rate because of Trump’s crackdown on Black Lives Matter protests? What about youth? What about the Republican or Democratic base? Voter turnout is a major strategic concern for candidates running for office. This concern was especially pronounced in 2020 given mask mandates and stay-at-home orders due to COVID-19. Elections can normally be won or lost by getting one’s supporters to the polls and keeping the opponent’s supporters at home. In 2020, the Biden campaign encouraged Americans to take advantage of the opportunity to use mail-in or early voting, whereas the Trump campaign raised questions about the security of votes that were not cast in person.

The activity surrounding get-out-the-vote (GOTV) efforts in any given election occurs against the backdrop of historically anemic turnout rates in the United States over the past fifty years. Turnout rates declined dramatically after 1960, leading many commentators to worry about the future of American democracy and many scholars to examine what was going on. Two major explanations were the focus of this research: institutional impediments to voting and individual-level attitudes that might increase or decrease turnout. We look at historical trends in voter turnout and at the institutional and attitudinal factors that affect whether people vote. We also address mobilization efforts by political parties and candidates and their impact on turnout.

While turnout is obviously important in a democracy, the American electorate can participate in many other ways that have an impact on elections. Americans can donate money to a campaign, put a sign in their yard or a bumper sticker on their car, make telephone calls on behalf of a candidate, go door-to-door canvassing for a candidate, write letters to the editor supporting a candidate, and so on. Donald Trump’s strategy of holding rallies across the country, even during the pandemic, and Joe Biden’s socially distanced approach point to the importance of campaign activists and enthusiasts. We therefore examine in this chapter not only who votes but who gets involved more actively in campaigns as well.

Learning objectives for Chapter 3 include:

· Understanding turnout trends and who the voters and nonvoters are

· Exploring the institutional impediments that those in power have used to keep especially Black people from voting

· Examining the psychological motivations that set voters apart from nonvoters

· Learning what campaigns do to increase turnout and who the activists are who get more involved in elections


One of the most persistent complaints about the recent American electoral system is its failure to achieve the high rates of voter turnout found in other countries and common in the United States in the nineteenth century. U.S. voter turnout was close to 80 percent before 1900; modern democracies around the world frequently record similarly high levels. Turnout in the United States over the past one hundred years, in contrast, has exceeded 60 percent only in presidential elections, and throughout most of the twentieth century it rarely did even that.

These unfavorable comparisons are somewhat misleading. The voting turnout rate is the percentage of the eligible population that votes in a particular election (the number of voters divided by the number of eligible voters). This seems straightforward, but it isn’t. As Michael McDonald and Samuel Popkin point out, most reports on voter turnout use the voting age population (VAP) as the denominator rather than the voting eligible population (VEP).1 Not all people who are of voting age are eligible to vote because of state laws restricting voting to, for example, U.S. citizens and people who fulfill residency requirements. The voting turnout rate is ideally calculated taking into account all state-level restrictions. When the denominator includes the VAP rather than the VEP, the turnout rate appears lower than it actually is because the denominator is inflated. On the flip side, in some states, Blacks, women, and eighteen-year-olds were given the right to vote before suffrage was extended to them nationwide by the Fifteenth, Nineteenth, and Twenty-Sixth Amendments to the U.S. Constitution. Since the Constitution originally left it up to the states to determine voter eligibility, states varied in whom they let vote. Not including these groups in the denominator when they were actually eligible to vote within their states makes the turnout rate in earlier years appear higher than it actually was because of a deflated denominator.

Determining the numerator in the turnout rate is also surprisingly difficult. The total number of ballots cast throughout the country is unknown; some states report the total vote only for particular races. For example, not included are those who went to the polls but skipped the presidential race or who inadvertently invalidated their ballots. This “undercount” of votes cast also reduces the estimate of turnout. Most often, the number of votes cast for the highest office on the ballot is used as the numerator, but this method could miss some votes.

McDonald, who runs the United States Elections Project, has attempted to calculate turnout more accurately by correcting both the numerator and the denominator of the official figures on a state-by-state basis starting in 1980. The turnout rate is slightly higher when total ballots cast, rather than total votes for the highest office, is used, but the difference is usually less than 1 percentage point. The most pronounced difference comes from using the VEP instead of the VAP, especially when looking at recent elections. The VEP-based turnout rate in the 1980s was about 2 percentage points higher than the VAP-based turnout rate. This difference increased to just under 5 percentage points beginning in 2004, largely due to both an increase in the number of noncitizens in the United States and an increase in the number of ineligible felons. In 1980, 3.5 percent of the U.S. population was made up of noncitizens, and just over 800,000 were ineligible felons. By 2020, 7.8 percent of the population was composed of noncitizens, and 3.29 million were ineligible felons.2 Rather than the official VAP highest office turnout rate of 62 percent in 2020, the VEP highest office turnout rate was 66.7 percent. Thus, turnout tends to be low in the United States compared to other established democracies, but it is not as low as official statistics suggest.

Despite the difficulties in estimating turnout, the data in Figure 3-1 show that dramatic shifts in the rate of voter turnout have occurred over time. During the nineteenth century, national turnout appears to have been extremely high—always more than 70 percent. The biggest drop in turnout occurred after 1896, especially in the South. The steep drop in the South from 1900 to 1916 is in part attributable to the restrictions placed on African American voting and to the increasing one-party domination of the region. In many Southern states, whoever won the Democratic primary won the general election, making turnout in the primary much more important than turnout in the general election.

Political maneuverings in the South, however, cannot explain the decline in turnout in the rest of the country that occurred at about the same time. While the Republican Party became dominant in the non-South, leading to less competition and therefore less interest in general elections, the Progressive Era reforms of the late 1800s and early 1900s likely affected turnout rates across the United States.3 Party organizations in the latter part of the 1800s, referred to as party machines, “delivered” or “voted” substantial numbers of votes via party loyalists casting multiple votes, “voting tombstones” (dead people), or vote buying. The decline in voter turnout in the early twentieth century coincides with the introduction of electoral reforms, including the introduction of the secret, or Australian, ballot and the imposition of a system of voter registration.4 Prior to the electoral reforms, voters were given distinctively colored ballots from their political party and openly placed them in the ballot box. The Australian ballot provided for secret voting and an official ballot with all candidates’ names appearing on it, thereby decreasing party control of voting. Without the color-coded ballots, the parties couldn’t know for whom people voted, which meant they couldn’t reward or punish people according to their vote. Voter registration requirements were another useful tool for combating corruption, by limiting the opportunity for fraudulent voting, but they also created an additional barrier to voting that had the effect of decreasing the turnout of less motivated potential voters. Turnout in national elections reached a low in the early 1920s and then fluctuated up and down before experiencing the recent uptick in the early years of the twenty-first century. The year 2020 saw an especially high turnout, reaching almost 67 percent. Turnout hadn’t been this high since 1904.

Great differences in turnout among the states are concealed within these national data. Rates of voting in the South, as shown in Figure 3-1, were consistently low until recently, when they nearly converged with northern turnout, but state variation is still considerable. For example, states with the lowest turnout in 2020 were Oklahoma at 55 percent and Arkansas at 56 percent. States with the highest turnout were Colorado at 76 percent and Minnesota at 80 percent.5 Variation in turnout is considerable not only from state to state but also from one type of election to another. Elections vary in the amount of interest and attention they generate in the electorate. As Figure 3-2 demonstrates, high-salience presidential elections draw higher turnout, whereas low-salience elections, such as off-year congressional elections, are characterized by turnout levels that are 10 to 20 percent lower. Even in a presidential election year, fewer people vote for a candidate in U.S. House races than for president, although in 2020 the percentages of votes for each of these two races based on VEP were close. Primaries and local elections elicit still lower turnout. Most of these differences in turnout can be accounted for by the lower visibility of nonpresidential elections; when less information about an election is available to voters, their interest is diminished.

A line graph of the estimated turnout in presidential elections in the Nation, the South, and the Non-South between 1860 and 2020.Description

Figure 3-1 Estimated Turnout in Presidential Elections in the Nation, the South, and the Non-South, 1860–2020

Sources: Curtis Gans, Voter Turnout in the United States, 1788–2009 (Washington, DC: CQ Press, 2010) for the data from 1860 to 2010; Michael McDonald, “Voter Turnout,” United States Elections Project, accessed June 30, 2021,

Note: The states included in the South were Alabama, Arkansas, Delaware, Florida, Georgia, Kentucky, Louisiana, Maryland, Mississippi, North Carolina, Oklahoma, South Carolina, Tennessee, Texas, Virginia, and West Virginia. All other states were counted as non-South. The data on Southern states from 1864 were limited to only four states.

A line graph of the estimated turnout in presidential and congressional elections between 1860 and 2020.Description

Figure 3-2 Estimated Turnout in Presidential and Congressional Elections, 1860–2020

Sources: Curtis Gans, Voter Turnout in the United States, 1788–2009 (Washington, DC: CQ Press, 2010) for the data from 1860 to 2010; Michael McDonald, “Voter Turnout,” United States Elections Project, accessed June 30, 2021,

Voters and Nonvoters

While turnout rates vary across time and across states, political scientists are pretty clear on the demographic characteristics of those who vote and of those who choose not to vote. Campaign staffs care a great deal about who the voters and nonvoters are as well because what matters is whether their party’s base is getting to the polls and the other party’s base is staying home. We look at the voter turnout rates of various demographic groups in 2020 and compare these results to previous elections.

Socioeconomic status is a key predictor of turnout. People who are better educated, wealthier, and in more professional occupations consistently turn out to vote at a higher rate than those from a lower socioeconomic status. When respondents were asked in the American National Election Studies (ANES) surveys from 1972 to 2016 if they voted in the presidential election, slightly over 91 percent of those with a college education or postgraduate degree said they voted compared to only about 69 percent of those with a high school education or less. In 2020, both groups had slightly higher reported turnout rates: 94 percent of those with a college degree and 76 percent of those without a college degree cast a ballot. Even in this year of heightened participation among the less educated, the college educated voted at a higher rate. Granted, self-reported turnout is always higher than actual turnout numbers, in part due to the social desirability problem of people not wanting to admit they did not vote when they know they should have.6 Research suggests that the better educated seem to be more affected by the social desirability bias than the less educated when it comes to voting and therefore are more likely to claim to have voted when they did not. For example, the well educated are much more likely to view voting as very much a duty (62 percent) compared to the less well educated (45 percent), who are more likely to believe voting is a choice. It is harder to admit to not doing one’s duty than it is to admit that one simply chose not to vote. Even taking into account exaggerated turnout numbers, education is highly related to voting for a variety of reasons, including having a better understanding of the voter registration process, having greater interest in and knowledge about politics, and being part of a more politically active social network.7 Not surprisingly, family income plays out in much the same way. In the past forty years, ANES data show that 86 percent of those in the top third of family income say they voted in the previous presidential election compared to just over 61 percent of those in the bottom third. The most recent presidential election had increases in both, the bottom third reaching 69 percent and the top third reaching 93 percent, but the pattern remains. Clearly, socioeconomic status matters in American elections.

The elections of 2020 fit the pattern of demographic shifts that have been taking place over the past seventy years. It used to be the case that men turned out to vote at a higher rate than women, sometimes by as much as 12 percentage points (as happened in 1956). As Figure 3-3 shows, this tendency reversed itself in 2004 when women began to vote at a higher rate than men, especially in 2008 when there was a 6-percentage-point difference in reported turnout between women and men. There has been little change in the difference in voting rates between men and women since 2012, about 3 percentage points, including in 2020.

Race has long been a key factor when discussing turnout. Whites discriminated against Blacks primarily but not exclusively in the South when it came to registering and voting. National turnout figures for Blacks consistently showed them voting at a much lower rate than whites because of these discriminatory practices that decreased their voting eligibility. When registration and voting laws that discriminated against Blacks were removed, the turnout rate among Black voters increased. Figure 3-4 shows that while Blacks closed the distance with whites after the Voting Rights Act passed in 1965, they did not surpass whites in turnout until 2008, when Barack Obama first ran as the Democratic nominee for president. Black Americans voted at a higher rate than whites in both 2012 and 2016, but their turnout rate dropped to below that of white people in 2020.

A line graph of the voter turnout by gender between 1952 and 2020.Description

Figure 3-3 Voter Turnout, by Gender, 1952–2020

Source: American National Election Studies, available at

The ethnic group that has voted at a level lower than whites and Blacks is Latinos and Latinas. They voted at a higher rate than Blacks in 1988 and at parity in 1996, but since then they have had a significantly lower turnout rate than both Blacks and whites. Part of the reason behind these lower turnout rates is voter eligibility. In the past, some states gave noncitizens the right to vote, but today only citizens are allowed to vote in federal and state elections. Immigrants who are not U.S. citizens might be asked in a survey if they voted, but they are not eligible to vote. Even taking eligibility into account, however, Latin@s vote at a lower rate than whites and Blacks, as can be seen in Figure 3-4. One potential explanation is the possible language barrier some Latin@s might experience. A more likely explanation is that campaigns have been slow or inconsistent in targeting Latino and Latina voters. As this population has grown in the United States and as their vote has become more critical to election outcomes, future campaigns will be smart to target Latin@s in a meaningful way.

A line graph of the voter turnout by race and ethnicity between 1952 and 2020.Description

Figure 3-4 Voter Turnout, by Race/Ethnicity, 1952–2020

Source: American National Election Studies, available at

One demographic group that consistently gets a lot of attention for not voting is young people. People in the eighteen to thirty-four age cohort consistently vote at a lower rate than older people, sometimes by 20 percentage points (see Figure 3-5). Even in a good year, such as 2004, youth turned out to vote significantly less than older people. The increase in reported turnout in 2004 and 2008 was most impressive among the youngest group of eighteen- to twenty-five-year-olds, whose turnout increased from 48 percent in 2000 to 64 percent in 2008 (data not shown). Many analysts wondered what would happen to youth turnout in 2016. Bernie Sanders garnered enthusiastic support from young people during the primaries, and his failure to get the Democratic Party’s nomination was frustrating to many of his supporters. Some Sanders supporters insisted they could not vote for Hillary Clinton, which raised the specter of a decline in voter turnout especially among young people. Figure 3-5 shows this was not the case. The turnout of young people increased slightly between 2012 and 2016, suggesting there was not a pronounced Sanders effect on young adult turnout. Those under thirty-five had their highest turnout since 1960 in the 2020 election, when 73 percent reported voting, an 11-percentage-point increase over 2016. Older people also increased their turnout in 2020 but not anywhere near as much as younger people.

A line graph of the voter turnout by age between 1952 and 2020.Description

Figure 3-5 Voter Turnout, by Age, 1952–2020

Source: American National Election Studies, available at

There are many reasons why young people are less likely to vote than older people, including motivational and institutional factors. Young people often incorrectly think that politics doesn’t have much to do with their lives when in reality it does. Aside from the direct connection with certain issues, such as student loan rates, many laws debated by Congress have a big impact on young people, such as health care reform, the spending of money on defense versus social programs, and so on. A contextual factor that likely affects youth turnout is the attention, or lack of attention, they receive from the candidates running for office. Candidates know that young people vote at a much lower rate than older people and therefore often tailor their messages to older people to capture their votes. When candidates take the time to talk specifically to younger people, as Obama did in 2008 and 2012, they are able to both increase youth turnout and gain a big share of their vote. Finally, much of the nonvoting among young people may be attributed to the unsettled circumstances of this age group rather than to simple disinterest in politics. Military service, being away at college, geographic mobility with the possible failure to meet residence requirements, and the additional hurdle of initial registration, along with receiving less attention from candidates, all create barriers to voting for young citizens that are less likely to affect older ones.


Political behavior, including whether people vote or not, takes place within a certain context. That context includes the institutional arrangements that make up the electoral system in the United States. Would turnout be higher if elections were held on a weekend instead of a Tuesday? Would it be higher if the government automatically registered its citizens instead of having citizens take the initiative to get registered? Would it be higher if the United States had more competitive electoral districts? People have jobs, take care of families, and attend school, all of which make it difficult at times to fit politics into their already busy lives. The institutional arrangements surrounding elections can make it easier or more difficult for people to get to the polls. In essence, the easier it is to vote, the more people will turn out to vote. Unfortunately, institutional impediments have frequently been used by whites to keep Blacks from voting.

Restrictions of Suffrage

Decisions about the institutional arrangements used in elections are inherently political and often partisan because they affect who can vote and how easily they can vote. After the Civil War, Republicans were eager to enfranchise Blacks, figuring that this new group of voters would vote Republican. In the early 1970s, Democrats were eager to enfranchise eighteen- to twenty-year-olds, figuring that they would vote Democratic. Reformers of all sorts encouraged the enfranchisement of women as a means of promoting their own goals. Many optimistically saw women voters as the cure for corruption in government, as unwavering opponents of alcohol consumption, and as champions of virtue in the electorate. Expansions of suffrage are quite rare, however, compared to attempts to restrict suffrage. The most notorious of these efforts was the effective disenfranchisement of Blacks in the South during the late nineteenth through the mid-twentieth centuries.

Several techniques for disenfranchising Blacks were used after Reconstruction in the South, and from time to time some of these techniques were applied in the North on a more limited basis to restrict the electoral participation of immigrants. The most common methods included white primaries, the poll tax, literacy tests, discriminatory administrative procedures, and intimidation. In some Southern states, only whites were allowed to vote in the party primary (the crucial election in one-party states), under the rationale that primaries to nominate candidates were internal functions of a private organization. In 1944, the U.S. Supreme Court ruled such white primaries unconstitutional on the ground that the selection of candidates for election is a public function in which discrimination on the basis of race is prohibited. The now-unconstitutional poll tax, whereby each individual was charged a flat fee as a prerequisite for registration to vote, was used for years and no doubt disenfranchised both poor Blacks and poor whites. The literacy test gave local officials a device that could be administered in a selective way to permit registration of whites while prohibiting that of Blacks. Registrars could ask Blacks to read and interpret the state constitution, for example, and insist they had not done a satisfactory job, whereas they might ask whites only to sign their names. To remain effective over long periods of time, these and other similar administrative devices probably depended on intimidation or the use of violence against Blacks.8 The outlawing of the poll tax through constitutional amendment and the suspension of literacy tests by the Voting Rights Act of 1965 and its extensions eliminated two important restrictions on the right to vote.

Felon disenfranchisement remains a major state restriction on suffrage, although there is great variation from state to state. In all but two states (Maine and Vermont) and the District of Columbia, prison inmates cannot vote. In many states, convicted felons cannot vote until they have served their entire sentence—in other words, served their time in prison and completed probation or parole. In some states, a felony conviction entails a permanent forfeiture of voting rights. With the prison population growing, this amounts to a sizable restriction of the franchise. A study by the Sentencing Project estimates that in 2016, over 6 million citizens were ineligible felons, compared to 3.34 million in 1996 and 1.17 million in 1976. This means that in 2016 approximately 2.5 percent of the VAP was disenfranchised because of a felony conviction. Over 10 percent of the VAP was disenfranchised in Florida, which had particularly strict laws until Florida voters passed an amendment in 2018 allowing most people who have served their sentences to vote. A legal battle ensued over whether all legal financial obligations have to be paid before people regain their right to vote.9

The number of felons who were disenfranchised dropped between 2016 and 2020 to 5.17 million, or 2.27 percent of the U.S. VEP. Blacks remain the group hardest hit by felon disenfranchisement laws because of their higher rates of incarceration and because they tend to live in states that disenfranchise felons for life, even after they have served all of their sentence. The Sentencing Project study estimates that 6.2 percent of Black adults are disenfranchised because of felony convictions (down from 7.4 percent in 2016), although certain states have much higher percentages. Over 20 percent of Blacks in Tennessee and Wyoming are disenfranchised. The disenfranchisement rate of non-Blacks is 1.7 percent, although this figure includes Latin@s who have a higher disenfranchisement rate than whites (over 2 percent). Women make up approximately one-fifth of the total disenfranchised population.10

Felon disenfranchisement clearly decreases the VEP, but does it decrease voter turnout? This question is more difficult to answer. On the one hand, if the people who are disenfranchised because of a felony conviction would not have voted anyway, then voter turnout is not affected by these state laws. Felons often come from certain demographic groups—primarily young people, the less educated, and the poor—that are less likely to vote. Based on these arguments, various researchers have found that felon disenfranchisement laws do not have a significant impact on turnout rates after taking into account several demographic factors.11 Other scholars, however, have estimated a much larger impact. By matching felons and nonfelons on such characteristics as gender, race, age, and education, Christopher Uggen and Jeff Manza estimate that just over a third (35 percent) of disenfranchised felons would have voted in presidential elections in recent years. They also estimate that a large proportion of these disenfranchised felons would vote for Democratic candidates.12 In states with higher percentages of disenfranchised felons and in close elections, these disenfranchised nonvoters could affect election outcomes. Putting felon disenfranchisement aside, some research shows that even being arrested (and not convicted) increases distrust in government and decreases attachment to the political system, leading to significantly lower turnout rates among those who have experienced the criminal justice system.13 Even if states do not have severe felon disenfranchisement laws, they likely have citizens who do not vote in part because of their experiences with the criminal justice system. Evidence suggests, however, that simple outreach campaigns providing information to felons who have served their sentences can increase the participation rate of this population.14

Reforms and Institutional Impediments to Voting

Historically, the United States has stood out as being less voter friendly than many Western democracies. In many of these countries, governments maintain registration lists instead of placing the burden of registration on the individual. In the United States, citizens in all states but one—North Dakota—must register to be able to vote. States vary dramatically in how many days prior to the election people must register, ranging from Election Day registration in such states as Minnesota and Colorado, where people can register immediately prior to casting their vote, to registering thirty days in advance of the election in such states as Texas and Ohio. West Virginia has a twenty-one-day deadline, Virginia a nineteen-day deadline, California a fifteen-day deadline, and Massachusetts a ten-day deadline. Not only must people register when they first vote; they must reregister each time they move. According to the U.S. Census Bureau, 9 to 11 percent of Americans move in any given year.15 College students are especially mobile. Regulations also typically cancel the registration of people who fail to vote in a few consecutive elections. With the various registration deadlines along with different rules concerning residency requirements set by each state, simply getting registered can appear daunting, and registration requirements raise the costs of political involvement, costs that a significant number of citizens choose not to assume.

Registering is much easier today than it was in the past, when people had to travel to the county seat to register, but this added step increases the costs of voting. Classic studies estimated that turnout in the United States would increase by 9 to 14 percent if people could register to vote on Election Day.16 Because of concern over low voter turnout and the role registration requirements likely play in that low turnout, reformers have worked hard to make registering easier. In 1993, Congress passed the “motor voter” bill, which provides that registration forms will be available at various governmental agencies that citizens visit for other purposes. These include agencies where motor vehicles are registered and driver’s licenses are obtained; however, because of a Republican-sponsored amendment, states are not required to provide them at unemployment and welfare offices. Because the unregistered tend to be poorer and less well educated, Democrats, who traditionally represent such groups, hoped (and Republicans feared) that reducing registration obstacles would increase the number of Democratic voters.

Many of the same political considerations were at play in the passage of the Help America Vote Act (HAVA) of 2002, which mandated that states provide provisional ballots for those citizens who believe they are registered to vote but whose names are not on the registration rolls. Congress passed HAVA in reaction to the voting debacle in Florida in the 2000 presidential election between Al Gore and George W. Bush. Aside from poorly designed butterfly ballots, hanging chads, and miscounted overseas absentee ballots,17 some voters were turned away from polling places because they did not appear on the voter lists even though they insisted they were registered. At a primarily Black precinct in Fort Lauderdale, election workers turned away about one hundred people who came to vote because the voter list indicated they had not registered.18 Similar scenes were played out across Florida. As noted, a provision in HAVA was designed to ensure that registered voters could vote by allowing people to cast a provisional ballot. If people show up to vote on Election Day and their name does not appear on the voter list, election workers must now offer them a provisional ballot. If they are registered, their vote will be counted. Much of the debate in Congress over this provision in HAVA involved the form of identification that voters seeking to cast provisional ballots would need to produce at the polls or within a certain time period after the election. Republican legislators sought more rigid standards to prevent voter fraud; Democratic legislators generally argued for keeping the barriers to a minimum.19 Despite Republican fears that easing registration requirements would bring more Democrats to the polls, the main impact of making it easier to register seems to have been a decline in the percentage of people not registered to vote and a slight increase in the proportion of the officially registered who do not vote (see Figure 3-6). Overall, the effect has been relatively small. Failure to register prohibits voting, but registering does not ensure turnout.

Registration requirements are not the only institutional impediment to voting. Traditionally, voting in national elections has been done on just one day, the Tuesday after the first Monday in November, at polling sites located in the precincts near where people live. These polling sites are open at a set time, in many states from 7 a.m. to 7 p.m., although the opening and closing times vary by state. Reformers have argued that having elections on a weekday, when many people work, decreases turnout. The more inconvenient voting is for people, the less likely they are to vote.

A line graph of the registration status and voter turnout between 1952 and 2016.Description

Figure 3-6 Registration Status and Voter Turnout, 1952–2016

Source: American National Election Studies, available at

To lessen the impact of these obstacles, reforms have opened up when and where people can vote by making it easier to vote through mail ballots, early voting, absentee ballots, and the use of voting centers. Some states mail ballots to registered voters and allow them to vote by mail. Oregon has pursued these possibilities most aggressively since 2004, sending the entire electorate mail-in ballots rather than having in-person voting. Four states (Colorado, Hawaii, Utah, and Washington) have moved to all-mail elections, following Oregon’s lead, although other voting options are available. Other states have loosened the conditions under which voters can request absentee ballots or have set up systems for in-person early voting in the weeks prior to the election, referred to as nontraditional voting. Only 11 percent of votes cast in 1996 were nontraditional votes. The percentage of nontraditional votes rose to 31 percent in 2008 and by 2016 was at 40 percent. Fears about how people would get to the polls during the pandemic led many states to open up early voting and mail ballot opportunities, and many people took advantage of these opportunities. In 2020, the percentage of nontraditional votes jumped to 69 percent.20 The largest increase came in the use of mail or absentee ballots, which jumped from just over 20 percent in 2016 to almost 50 percent of all votes cast in 2020.21

If making voting more convenient is the answer to higher turnout, then the series of reforms put in place in recent years should lead to a significant increase in votes cast. The act of voting is much more convenient than it used to be, thereby lowering the costs associated with voting. Research on the impact of these reforms, however, has not shown the effects many reformers had hoped would occur. The people taking advantage of these reforms tend to be those who would vote anyway; it is just easier for them now. And these reforms have not increased substantially the turnout of underrepresented groups, such as people of color or young people.22 It can’t be denied, though, that turnout has been increasing since the mid-1990s over the time that voting has become more convenient.

In response to the increased access to voting, some states and localities have taken steps to make voting potentially less convenient. Many of these efforts occurred long before the 2020 elections, including the closing of polling places and the requirement instituted by some states that voters provide an official form of identification to be able to vote. In 2013, the Supreme Court struck down, in Shelby County v. Holder, the restrictions in Section 5 of the 1965 Voting Rights Act that required states with a history of voting discrimination to get federal court permission to make changes in how their elections are carried out. Since that decision, many counties, especially in the South, have closed voting stations. Between 2012 and 2018, over 1,600 polling places were closed, including 1 in 10 in Texas and 1 in 5 in Arizona. Almost 1,200 of these occurred after the 2014 elections.23 The argument for closing these voting stations was that they were underutilized and it was therefore cost efficient to close them. Opponents of the closures argued that having fewer polling sites would make voting less convenient and would increase the time needed to vote.24 The availability of voting stations and ease of access to them vary for people of different racial backgrounds. One study of Wisconsin voters in 2018 showed that Latin@ voters systematically faced longer waits in line than other voters and Black voters had longer commutes to their polling places.25

Instituting voter identification (ID) laws has also increased. After the 2000 presidential election, states began passing voter ID laws in droves. According to the National Conference of State Legislatures, thirty-six states have voter ID laws in place (although North Carolina’s law had not gone into effect by the end of 2020 due to court cases in the works).26 These state laws vary, with some states requiring an official photo ID, others requiring a nonphoto ID, and still others requesting but not requiring an ID. What happens if people do not have an appropriate form of identification also varies by state. In some states, people who do not have an appropriate ID can vote using a provisional ballot but must bring their ID to election officials within a few days of the election to have their vote counted. Other states allow those without an acceptable ID to vote if they sign a voucher attesting to their identity or if a poll worker vouches for them.

A University of Massachusetts Amherst/WCVB poll released in April 202127 found a large majority of Americans, 67 percent, favor the requirement that voters show a photo ID to be able to vote. Partisanship played a role in this support, with 94 percent of Republicans, 71 percent of independents, and 45 percent of Democrats supporting a voter ID requirement. Supporters of voter ID laws are more likely to worry about voter fraud, arguing that people who are ineligible to vote are casting ballots and potentially swaying election outcomes, whereas opponents are more likely to worry about voter suppression, arguing that people who are eligible to vote but cannot afford a photo ID are being turned away from the polls. Not surprisingly, Republicans are more likely than Democrats to believe that preventing election fraud is a priority even if it makes voting harder (88 percent of Republicans vs. 22 percent of Democrats). Democrats, in contrast, are much more likely than Republicans to support automatic voter registration (84 percent to 24 percent), vote by mail (90 percent to 23 percent), and automatic absentee ballots (76 percent to 13 percent).

Little evidence exists that there is actual widespread voter fraud, especially of the type that a voter ID law would presumably stop. When Indiana passed its photo ID law among claims of widespread voter fraud, no cases of in-person fraud had been prosecuted, and a special investigation over almost two years of voter fraud in Texas led to only thirteen indictments, six of which involved people helping friends with a mail-in ballot.28 Fears of widespread voter fraud appear to be misplaced. One study found that Americans were more likely to claim that they had been abducted by an alien than to have impersonated another voter.29 It could be the case, however, that regardless of the reality of voter fraud, having voter ID laws could increase confidence in the election process. This argument was made by the Supreme Court in its Crawford v. Marion County Election Board (2008) decision. Whether fraud exists or not, increasing confidence in elections might be an important outcome and a reason to pass voter ID laws. Research shows, however, that people who live in states with strict voter ID laws are no more confident in election outcomes, or no less likely to believe that fraud exists, than people who live in states with no voter ID laws.30

Evidence on the effects of voter ID laws on turnout has been mixed, although any effects appear to be small. Some researchers have found a moderate impact on turnout among people who are the least likely to have the type of identification required in the state, especially people of color. Others, however, present evidence for minuscule effects.31 Stephen Ansolabehere found that less than two-tenths of 1 percent (seven people out of four thousand) did not vote in 2008 because of voter ID problems. It is certainly possible that the most strict voter ID laws, which have only more recently been passed, could have a detrimental impact on turnout, a potential that will need to be watched. For example, research shows that Hispanics and Blacks are much more likely to be asked by poll workers to show some type of identification, especially photo IDs, even when there is no state law in place requiring identification to vote.32 The unequal treatment of voters at the polls based on race raises serious concerns about voter ID laws.

After the 2020 election, states have ramped up their efforts to put more restrictions on voting. According to the Brennan Center for Justice, legislators in forty-seven states have introduced bills to restrict voting in one way or another. By May 2021, five states had passed restrictive legislation. For many states, the bills were still moving through the legislative process. In response, legislators have proposed codifying in law the more expansive practices of 2020, and the U.S. House of Representatives passed H.R. 1 in March 2021 to expand voter registration and increase voter access.33 The bill at the time of this writing sits in the U.S. Senate. The current fights over institutional impediments to voting matter because they affect voter turnout, and who votes matters in a democracy.


Election processes and rules clearly play a role in encouraging or dampening voter turnout. Any obstacles increase the cost of voting, and even the smallest cost can cause people to find other things to do on Election Day. But many people vote even when obstacles and costs exist, just as many people do not vote even when voting is relatively easy. Making the effort to register, to become informed about an upcoming election, and to cast a ballot demands a certain level of interest and engagement in politics that cannot be created simply by reducing institutional obstacles.

The people most likely to vote, especially in midterm or local elections, are highly interested in politics, are more likely to be knowledgeable about politics, and feel strongly attached to a political party, largely because strong identifiers care more about who wins.34 Even though interest in politics is strongly correlated with voting, about half of those who say they have hardly any interest (not very and not at all interested) do vote in presidential elections (and in 2020, a whopping 72 percent of the uninterested said they voted), suggesting that other factors are also at work. One of these is a sense of civic duty—the attitude that a good citizen has an obligation to vote. Many Americans see voting as an obligation of citizenship (51 percent see it more as a duty than a choice), and when they vote, they feel a sense of gratification that overrides any cost of voting they might incur.35 When voting is viewed as a civic norm, people with a strong sense of civic duty vote because of the intrinsic satisfaction they get from doing what they know is right, or they vote because of extrinsic pressure to conform to a social norm. In a clever experiment, Alan Gerber, Donald Green, and Christopher Larimer found that when people were reminded of the obligation to vote, turnout increased by 1.8 percentage points compared to people who received no message. The biggest impact on turnout, however, came from people who were told that their neighbors would know whether they voted or not. In this case, turnout increased by over 8 percentage points.36 A similar effect was found in a study of Facebook users that manipulated whether people saw posts in their Facebook feed that showed their friends voting.37 Civic duty increases turnout not just because people feel good when they have done what they know they ought to do; it also increases turnout because people experience social pressure to vote. It should come as no surprise that polling places often hand out “I voted” stickers, allowing voters to publicly display the fact that they fulfilled their civic obligation to vote.

Political scientists have begun to wonder more recently if there are even deeper explanations for people’s voting behavior, deeper in the sense that there might be a genetic component that leads people to vote or not to vote. Testing genetic influences is not easy, given that there is obviously not a gene for voting, but scholars have been able to use studies of twins to compare monozygotic twins (popularly known as identical twins, who share 100 percent of their genes) and dizygotic twins (popularly known as fraternal twins, who share only 50 percent of their genes, which is true for all biological siblings). We know that parents who vote are much more likely to have children who grow up to vote, and this is likely affected somewhat more by genetic influences than socialization influences. James Fowler and his colleagues found that over 50 percent of people’s turnout behavior can be explained by genetic heritability, whereas about 35 percent can be explained by shared environment. Environment matters, but genes play a big role. The influence of genetics on voting might well occur through the large role they play in explaining attitudes related to turnout, such as interest, partisanship, civic duty, and political efficacy.38


Our discussion of voter turnout thus far has implied that there are people who vote (sharing certain demographic characteristics, such as higher education level and income; certain attitudes, such as interest and civic duty; and certain genetics) and people who do not vote (those who do not share these characteristics). This is clearly not the case. Some people never vote, estimated to make up about 10 percent of eligible voters, and they are unlikely to go to the polls regardless of what institutional rules are in place or what is done to get them to vote. Constant voters, about 25 percent of eligible voters, do not need to be prompted to go to the polls, and they will overcome whatever obstacles might get in their way. They always vote. It is the remaining 65 percent who are the intermittent voters. They are more likely to vote in high-salience elections (e.g., presidential elections) and when voting is convenient. Since reforms have, in general, made voting easier, the trick now is to get these intermittent voters to vote regularly. Getting these voters to the polls is a target of mobilization efforts.39

Millions of dollars are spent by campaigns, partisan groups, and nonpartisan organizations to get people to the polls on Election Day. These GOTV efforts include door-to-door canvassing, leaflets, door hangers, direct mail, email, and phone calls. In a series of field experiments, political scientists have tested the varying effects of these GOTV strategies to determine which lead to higher turnout.40 Door-to-door canvassing, where campaign volunteers ring doorbells and ask the targeted individuals to be sure to vote, tends to be more effective than using the phone or mail. Personal, face-to-face requests elicit greater compliance than impersonal requests, but people also tend to follow through when they have made a commitment publicly.41 People like to think of themselves as consistent, and the only way to be consistent after telling a canvasser that they will vote in the upcoming election is to vote. Given this logic, it makes sense that phone calls can be effective as well if there is a more personal touch involved—for example, when the call is made by volunteers or when professional phone banks use a more interactive and conversational approach rather than robotic calls.42 Major advances in GOTV efforts came into play in the 2012 presidential election with the use of “big data” and microtargeting. Both the Barack Obama and Mitt Romney campaigns bought demographic data from companies that gather personal data on everything from shopping habits to financial problems. They gathered online data themselves on such things as social networks. As New York Times reporter Charles Duhigg wrote before the election,

They have access to information about the personal lives of voters at a scale never before imagined. And they are using that data to try to influence voting habits—in effect, to train voters to go to the polls through subtle cues, rewards, and threats in a manner akin to the marketing efforts of credit card companies and big-box retailers.43

The campaigns used the information to contact potential voters and apply targeted pressure to get them to vote, although appearing to know too much personal information can backfire by seeming creepy.

The Obama campaign was especially advanced in microtargeting. Jim Messina, Obama’s campaign manager in 2012, set up campaign headquarters in Illinois and hired sixty data analysts to analyze all the data needed to microtarget. Being able to microtarget gave the Obama campaign a definite edge over the Romney campaign. In a story told by Messina, Obama volunteers were canvassing a neighborhood at the same time as Romney volunteers. The Romney volunteers knocked on every door on one side of the street, finding that half of the people were not home and the other half were Obama supporters. The Obama volunteers were told to knock on only two doors and to speak with certain people at those houses, people who fit the profile of being potential Obama supporters and who could be nudged to get to the polls on Election Day. Because of the analysis of big data, Messina said, the Obama campaign was able to target the houses of undecided voters who had a good probability of voting, and they were able to tell their volunteer canvassers what to say. The Obama volunteers knocked on the relevant doors, talked to the relevant people, and were able to move on to the next neighborhood, while the Romney volunteers were still knocking on doors that would not elicit Romney voters.44 The Obama campaign even ran experiments on volunteer phone calls to potential supporters to test whether it was more effective to control the message in the calls or let the voters talk about issues of their own choosing; the campaign persuaded more people to support Obama when sticking to the campaign script than when letting the voters lead the discussion.

GOTV efforts were decidedly skewed in 2016. While Hillary Clinton’s campaign had a less robust ground game than Obama’s did in 2012, she had more than twice as many field offices as Trump. Field offices are places where people can go to volunteer their services for the campaign. They therefore are the hub from which volunteers go out into neighborhoods to knock on doors and from which phone calls are made to energize voters. Trump’s strategy was to piggyback on the efforts of the Republican National Committee, which was a more efficient approach but less under the control of the candidate.45 The Trump campaign’s emphasis was on utilizing untraditional methods to increase enthusiasm and the desire to vote, including his signature rallies that generated significant media coverage.

The campaign strategies of the top two contenders in 2020 could not have been more different. In the midst of the pandemic, with mask mandates and stay-at-home orders in effect in many parts of the country, Democrat Joe Biden chose to hold occasional small events with strict social distancing, but he focused more on holding virtual gatherings and airing television ads. Rather than going door-to-door to mobilize voters, the Biden campaign staff used mobile offices to try to generate enthusiasm and engagement. Republican Donald Trump, in contrast, held large rallies at which his supporters wearing “Make America Great Again” caps gathered to cheer their candidate on. Trump traveled far and wide across the country and often chastised his opponent for hiding in his basement while he was out meeting with voters.46 Turnout was historically high in 2020, likely not because of the differing GOTV efforts of the two candidates but because people cared so much about which candidate won. Most Democrats (91 percent) and almost as many Republicans (84.5 percent) said they cared a lot or a great deal about who won the election.

Even if voter mobilization efforts were highly successful, turnout rates would not be consistently high across all elections. Context still matters. Good candidates, competitive races, and high-salience elections are more likely to bring out voters than are other, less invigorating races.47 In low-salience races, such as a local race for mayor or even a midterm congressional election, mobilization efforts are aimed at intermittent voters who vote frequently but not always and just need a nudge. GOTV campaigns can also be successful by targeting probable voters who are undecided. The problem is that there have been relatively few undecided voters in most recent elections. Defining undecided voters as those who stated in a survey that they were undecided and did not say they were leaning toward one candidate or the other when pushed, Larry Bartels and Lynn Vavreck found that only 5 percent of survey respondents in 2012 were actually undecided in the months before the election. These undecided voters not surprisingly tended to identify as independents and moderates, were not terribly knowledgeable about politics, and tended not to follow political news much.48 While people might not be able to make a conscious decision, however, they often have implicit leanings toward one candidate or the other, and this unconscious preference is a good predictor of who they will actually vote for on Election Day.49 In 2016, the number of people who claimed to be undecided was higher than normal (10 percent of the ANES respondents). They also turned out to vote in unusually high numbers, and they tended to vote for Trump.50 By 2020, the undecideds had dropped to the more normal 5 percent.

Who Participates in Campaigns?

Getting out the vote is what candidates need to do to win, but campaigns would be hard-pressed to get out the vote without a large amount of unpaid help. Candidates and political parties rely heavily on volunteers during the election season to do the canvassing, the stuffing of envelopes, and the calling of potential voters. They hope their supporters get out the word on their candidate by putting up yard signs or placing bumper stickers on their cars or talking to friends, relatives, and coworkers to drum up support for the candidate. And with the tremendous cost of campaigns, especially in recent years, campaign organizations eagerly solicit supporters’ donations. Voting is a relatively easy way to participate in politics compared to the initiative and costs (in time and, for donations, money) associated with other types of campaign activity. Campaign activists have to be highly motivated both to figure out what they need to do to be involved and to participate in the activity. Most Americans are not motivated to be activists. A focus group participant summed it up nicely when she said, “When I leave here [the focus group discussion], when I walk out this door, I’m not going to volunteer for anything. I’m not going to get involved in anything. I mean I know this. I’m not going to pretend I’m some political activist. I’m lazy. I’m not going to do it. I’m too busy obsessing on other things going on in my life.”51 It comes as no surprise that fewer people are involved in campaign activities than vote.

Figure 3-7 shows Americans’ involvement in various campaign activities over time. The first thing that stands out is that Americans are much more likely to try to influence other people in how they should vote than to be involved in other campaign activities. ANES respondents were asked, “During the campaign, did you talk to any people and try to show them why they should vote for or against one of the parties or candidates?” On average, about a third of the respondents said yes, they did try to influence people’s votes. This percentage skyrocketed to just under 50 percent in the 2004 election and dropped to 43 percent in 2008. The percentage of those trying to influence other people’s votes has remained at this level over the recent election cycles, between 40 and 44 percent, but this level was still higher than the average over the time period covered by ANES. Making the effort to try to persuade people how to vote indicates a strong interest in the outcome of the election and enough knowledge about the campaign to be able to make an argument on behalf of a party or candidate. It is interesting to note that while voters are the most likely to try to influence others, people who end up not voting do so as well. Over the past several elections, approximately 50 percent of voters and just over 20 percent of nonvoters tried to influence others’ vote choice. These percentages were 42 percent of voters and 26 percent of nonvoters in 2020. We can’t know what the nonvoters had to say while trying to persuade other people, but we do know that in the past election, half (51 percent) of nonvoters claimed to “care a great deal” or “a lot” about which party won the election. Granted, voters were more likely to care (87 percent), but it is clear that a lot of nonvoters care enough about the outcome of the election to want to persuade others for whom to vote. The other campaign activities are not as popular, with only 3 to 5 percent of Americans working on campaigns and just under 10 percent attending political meetings. With the pandemic limiting social gatherings in 2020, attendance at meetings, rallies, speeches, and dinners dropped to just over 5 percent. Many more people, 12 percent, said they had attended online meetings, rallies, speeches, and dinners, a reflection of the limits on social gatherings during the pandemic.

The second aspect of Figure 3-7 that stands out is that campaign activities overall increased significantly in 2004 but by 2012 had dropped to more normal levels (except for trying to influence people). People’s willingness to display buttons or stickers and to donate money increased in 2020, but 2004 still stands out as a particularly engaged election. So what is special about 2004? Much of the commentary leading up to the 2004 election focused on energized youth and upset Democrats who were still smarting after the 2000 election and the debacle in Florida. The closeness of the 2000 election led Democrats to emphasize the need for their followers to vote. The stepped-up efforts of Rock the Vote and Sean P. Diddy Combs’s “Vote or Die,” both aimed at increasing youth turnout, and the large number of Democrats thinking that year’s presidential election was “the most important of their lifetimes” led to speculation that youth and Democrats would be much more engaged in 2004 than was usual.52 The uptick in campaign activities in 2020 is likely due to a similar dynamic: Democrats were upset about winning the popular vote but losing the Electoral College vote in 2016, especially given their dislike of Trump (more on this in Chapter 5). Youth were highly energized by the Black Lives Matter protests over the summer of 2020, a topic we address in the next chapter.

A line graph of different campaign participation between 1952 and 2020.Description

Figure 3-7 Campaign Participation, 1952–2020

Source: American National Election Studies, available at

We have raised a number of different aspects of campaigns and voting that could influence whether people get engaged in campaigns. We are now going to put much of it together to analyze who is most likely to be engaged in campaigns, from voting to donating. Looking at certain factors, such as age or income, tells us a lot, but sometimes these variables overlap. For example, older people are more likely to have higher incomes than younger people. When looking at the relationship between age and participation in politics, is it age that explains what is going on? Or is it income level that is driving participation, and it just happens that older people have more income? To figure out the unique effects of each of the factors we have discussed, we need to run a regression analysis. The results tell us how much of an effect each variable has on campaign participation, controlling the effects of all of the other variables. Using this method, we can see whether, for example, age matters (controlling all else) and whether income matters (controlling all else).

We created a Campaign Activities Scale using ANES data that includes whether people voted, tried to influence others’ vote, worked on a campaign, attended a meeting in person, attended a meeting online, displayed a button or sticker, or donated money. A score of 0 means they did none of these (12 percent of the sample), and a score of 7 means they did all of these (only 0.4 percent had a score of 7). The mean score was 1.7, which means that the average person did just under two of these activities. To explain campaign activities behavior, we included as independent variables demographic information (sex, race, age, education, and income), interest in the election (whether the respondent cares about who wins the presidency and partisan strength, because strong partisans tend to care more about winning elections than those with a weaker or no partisanship), GOTV contact (whether anyone talked to the respondent about registering or voting), and believing voting is a duty (vs. thinking it is just a choice people can make). Table 3-1 shows the results of the regression analysis.

Table 3-1

Source: 2020 American National Election Studies, available at

The results are easy to interpret. All of the independent variables were scaled to range from 0 to 1. For example, an eighteen-year-old is coded 0, and someone who is 80 or older is coded 1. Everyone else falls somewhere between these two extremes. For race, we coded whites as 1 and people of color as 0. If the regression coefficient (B) is positive and the p value is smaller than .05, then the variable significantly increases campaign activity. If B is negative and the p value is smaller than .05, then the variable significantly decreases campaign activity. How big B is tells us how much a unit increase in the independent variable has an impact on campaign activities. For example, sex is coded 1 for males and 0 for females. The B is positive but extremely small (.011), and the p value is way above .05, which means that men and women are basically the same when it comes to campaign activity. In contrast, whites are more likely to participate in campaign activities than people of color, although whites only do .23 more campaign activities than people of color. (Keep in mind that the campaign activities scale ranges from 0 to 7, so an increase of .23 is very small.) Among the other demographic variables, older people are more likely to participate than younger people, the better educated are more likely to participate than the less educated, and wealthier people are more likely to participate than poorer people, all else equal.

The political motivation variables also explain involvement in campaign activities. The stronger people’s partisanship, the more likely they are to participate (by almost half of an activity). Being encouraged to vote by someone and believing voting is a duty have similar significant impacts on campaign activity. The best predictor of campaign participation, though, is caring who wins. The more people cared if Trump or Biden won, the more they participated in the campaign. Moving from not caring at all to caring a great deal moved people up the campaign activities scale by more than one activity. This result is hardly surprising, but it does suggest that holding everything else constant, if we can increase how much people care about who wins, participation in campaigns will increase quite dramatically. The downside, of course, is that if people care a great deal about who wins, then losing the election becomes much more difficult to accept.

Electoral context matters. What happens in the political world—being upset about previous election outcomes, having GOTV efforts target certain groups of people, caring passionately about who wins—can influence people’s behavior, making them more or less active in any given election. A common theme among democratic theorists is a desire for people to get more involved in politics. People need to vote, be informed about politics, participate in campaigns, and join organizations. The data throughout this chapter show that while many Americans achieve this standard of good citizenship, a great many do not. Political scientists have made significant advances in ascertaining what causes low participation rates and what can be done to increase citizens’ engagement in the political system. Having people care a great deal about politics because of what they see happening in Washington and around the country matters.

Study Questions

1. Voter turnout has gone up and down over time. Who are the voters, and who are the nonvoters?

2. How have institutional restrictions on voting been used to affect political outcomes, especially when it comes to restricting Blacks’ right to vote?

3. What are the psychological motivations that increase turnout?

4. What do campaign activists do, and who are they?

Suggested Readings

Green, Donald P., and Alan S. Gerber. Get Out the Vote! How to Increase Voter Turnout. Washington, DC: Brookings Institution Press, 2015. Recommendations on how to mobilize voters and increase turnout based on rigorous research.

Holbein, John B., and Sunshine Hillygus. Making Young Voters: Converting Civic Attitudes Into Civic Action. New York: Cambridge University Press, 2020. Combining psychological, economic, and child development theories, this book shows how education and civic reforms can increase younger voters’ turnout.

Leighley, Jan E., and Jonathan Nagler. Who Votes Now? Demographics, Issues, Inequality, and Turnout in the United States. Princeton, NJ: Princeton University Press, 2013. In-depth, thorough analysis of the characteristics of voters versus nonvoters.

Manza, Jeff, and Christopher Uggen. Locked Out: Disenfranchisement and American Democracy. New York: Oxford University Press, 2008. An indispensable examination of felon disenfranchisement laws and their impact on American elections.

McDonald, Michael P., and Samuel L. Popkin. “The Myth of the Vanishing Voter.” American Political Science Review 95, no. 4 (December 2001): 963–974.

Patterson, Thomas E. The Vanishing Voter: Public Involvement in an Age of Uncertainty. New York: Alfred A. Knopf, 2002. An analysis of political participation based on a huge, yearlong survey in 2000.

Rusk, Jerrold D., and John J. Stucker. “The Effect of the Southern System of Election Laws on Voting Participation.” In The History of American Electoral Behavior, edited by Joel Silbey, Allan Bogue, and William Flanigan. Princeton, NJ: Princeton University Press, 1978. A sophisticated analysis of the disenfranchisement of voters in the South in the nineteenth century.

Teixeira, Ruy A. The Disappearing American Voter. Washington, DC: Brookings Institution Press, 1992. A sophisticated and thorough analysis of the factors that have contributed to the decline in turnout in the United States and a discussion of the impact of proposed reforms.

Internet Resources

An important website for the analysis of aggregate turnout data is the United States Elections Project, Michael McDonald, the host of this website, uses strategies for reducing the error in turnout estimates and offers commentary on turnout.

The website of the American National Election Studies,, offers data on turnout in both presidential and off-year elections since 1952. In the Resources menu, click on the link for tables and graphs under “The ANES Guide to Public Opinion and Electoral Behavior” and then scroll down to “Political Involvement and Participation in Politics.” You also can examine turnout of numerous social groups from 1952 to the present.

Turnout and registration data for the nation and the states are available at the U.S. Census Bureau website, Click on “Browse by Topic” in the menu bar, then click on “Public Sector,” and then click on “Voting and Registration.”


1. Michael P. McDonald and Samuel L. Popkin, “The Myth of the Vanishing Voter,” American Political Science Review 95, no. 4 (2001): 963–974.

2. See United States Elections Project, accessed June 29, 2021,; and McDonald and Popkin, “Myth of the Vanishing Voter.”

3. On the one-party dominance argument, see E. E. Schattschneider, The Semisovereign People (New York: Holt, Rinehart, and Winston, 1960), especially Chapter 5; and Walter Dean Burnham, “The Changing Shape of the American Political Universe,” American Political Science Review 59 (March 1965): 7–28. On electoral manipulations and the subsequent electoral reforms, see Philip E. Converse, “Change in the American Electorate,” in The Human Meaning of Social Change, eds. Angus Campbell and Philip E. Converse (New York: Russell Sage Foundation, 1972), 263–337. For an analysis that alters the estimates of turnout, see Ray M. Shortridge, “Estimating Voter Participation,” in Analyzing Electoral History, eds. Jerome M. Clubb, William H. Flanigan, and Nancy H. Zingale (Beverly Hills, CA: SAGE, 1981), 137–152.

4. Jerrold D. Rusk, “The Effect of the Australian Ballot Reform on Split-Ticket Voting: 1876–1908,” American Political Science Review 64 (December 1970): 1220–1238.

5. See “Voter Turnout,” United States Elections Project, accessed June 29, 2021,

6. Allyson L. Holbrook and Jon A. Krosnick, “Social Desirability Bias in Voter Turnout Reports,” Public Opinion Quarterly 74 (Spring 2010): 37–67.

7. Brian D. Silver, Barbara A. Anderson, and Paul R. Abramson, “Who Overreports Voting?,” American Political Science Review 80 (June 1986): 613–624; Rachel Milstein Sondheimer and Donald P. Green, “Using Experiments to Estimate the Effects of Education on Voter Turnout,” American Journal of Political Science 54 (January 2010): 174–189.

8. For a treatment of these and many additional topics, see J. Morgan Kousser, The Shaping of Southern Politics (New Haven, CT: Yale University Press, 1974). See also Jerrold D. Rusk and John J. Stucker, “The Effect of the Southern System of Election Laws on Voting Participation,” in The History of American Electoral Behavior, eds. Joel Silbey, Allan Bogue, and William Flanigan (Princeton, NJ: Princeton University Press, 1978).

9. Christopher Uggen, Ryan Larson, Sarah Shannon, and Arleth Pulido-Nava, “Locked Out 2020: Estimates of People Denied Voting Rights Due to a Felony Conviction,” Sentencing Project, October 30, 2020,

10. Ibid.

11. Randi Hjalmarsson and Mark Lopez, “The Voting Behavior of Young Disenfranchised Felons: Would They Vote If They Could?,” American Law and Economics Review 12 (2010): 356–393; Thomas J. Miles, “Felon Disenfranchisement and Voter Turnout,” Journal of Legal Studies 33 (January 2004): 85–129.

12. Christopher Uggen and Jeff Manza, “Democratic Contraction? Political Consequences of Felon Disenfranchisement in the United States,” American Sociological Review 67 (December 2002): 777–803, p. 786.

13. Vesla M. Weaver and Amy E. Lerman, “Political Consequences of the Carceral State,” American Political Science Review 104 (November 2010): 817–833, p. 818.

14. Alan S. Gerber, Gregory A. Huber, Marc Meredith, Daniel R. Biggers, and David J. Hendry, “Can Incarcerated Felons Be (Re)integrated Into the Political System? Results From a Field Experiment,” American Journal of Political Science 59 (October 2015): 912–926.

15. “CPS Historical Migration/Geographic Mobility Tables” (Table A-1 Annual Geographic Mobility Rates, by Type of Movement: 1948–2020), U.S. Census Bureau, U.S. Department of Commerce, December 2020,

16. Raymond E. Wolfinger and Steven J. Rosenstone, Who Votes? (New Haven, CT: Yale University Press, 1980); G. Bingham Powell, “American Voter Turnout in Comparative Perspective,” American Political Science Review 80 (March 1986): 17–43.

17. Henry E. Brady, Michael C. Herron, Walter R. Mebane Jr., and Jasjeet Singh Sekhon, “‘Law and Data’: The Butterfly Ballot Episode,” PS: Political Science and Politics 34 (March 2001): 59–69; David Barstow and Don Van Natta Jr., “Examining the Vote; How Bush Took Florida: Mining the Overseas Absentee Vote,” New York Times, July 15, 2001.

18. Mireya Navarro and Somini Sengupta, “Contesting the Vote: Black Voters; Arriving at Florida Voting Places, Some Blacks Found Frustration,” New York Times, November 30, 2000.

19. CQ Almanac Plus 2002 (Washington, DC: Congressional Quarterly, 2003), 143.

20. Zachary Scherer, “Majority of Voters Used Nontraditional Methods to Cast Ballots in 2020,” U.S. Census Bureau, April 29, 2021,

21. “Voting by Mail and Absentee Voting,” MIT Election Data + Science Lab, accessed May 7, 2021,

22. See, for example, Adam Berinsky, “The Perverse Consequences of Electoral Reform in the United States,” American Politics Research 33 (July 2005): 471–491; Jeffrey A. Karp and Susan A. Banducci, “Going Postal: How All-Mail Elections Influence Turnout,” Political Behavior 22, no. 3 (2000): 223–239; Robert M. Stein and Greg Vonnahme, “Voting at Non-precinct Polling Places: A Review and Research Agenda,” Election Law Journal 10 (October 2011): 307–311.

23. Andy Sullivan, “Southern U.S. States Have Closed 1,200 Polling Places in Recent Years: Rights Group,” Reuters, September 9, 2019,

24. Elena Mejia Lutz, “Report: Texas Has Closed Most Polling Places Since Court Ruling,” Texas Tribune (Austin), November 4, 2016,; John Whitesides, “Polling Places Become Battleground in U.S. Voting Rights Fight,” Reuters, September 16, 2016,

25. Jordan Foley, Michael W. Wagner, Ceri Hughes, Jiyoun Suk, Katherine J. Cramer, Lewis A. Friedland, and Dhavan V. Shah, “Free and Fair? The Differential Experiences of Voting Barriers and Voting Policies in American Midterm Elections,” International Journal of Public Opinion Research (Forthcoming).

26. These numbers are accurate as of March 2021. “Voter Identification Requirements/Voter ID Laws,” National Conference of State Legislatures, May 25, 2021,

27. UMassAmherst UMass Poll, Toplines and Crosstabs April 2021 Election Reform, April 28, 2021, accessed May 7, 2021,*1yrvk7t*_ga*MjczNjg5NjIzLjE2MjA0MTQ0NzE.*_ga_21RLS0L7EB*MTYyMDQxNDQ3MC4xLjAuMTYyMDQxNDQ3MC4w&_ga=2.2687155.222364461.1620414471-273689623.1620414471.

28. Chandler Davidson, “The Historical Context of Voter Photo-ID Laws,” PS: Political Science & Politics 42 (January 2009): 93–96.

29. John S. Ahlquist, Kenneth R. Mayer, and Simon Jackman, “Alien Abduction and Voter Impersonation in the 2012 U.S. General Election: Evidence From a Survey List Experiment,” Election Law Journal: Rules, Politics and Policy 13, no. 4 (2014): 460–475.

30. Stephen Ansolabehere and Nathaniel Persily, “Vote Fraud in the Eye of the Beholder: The Role of Public Opinion in the Challenge to Voter Identification Requirements,” Harvard Law Review 121 (2008): 1737–1774; “Voter Identification,” MIT Election Data + Science Lab, accessed June 29, 2021,

31. Benjamin Highton, “Voter Identification Laws and Turnout in the United States,” Annual Review of Political Science (2017): 149–167; R. Michael Alvarez, Delia Bailey, and Jonathan N. Katz, “The Effect of Voter Identification Laws on Turnout” (California Institute of Technology Social Science Working Paper No. 1267R), January 17, 2008. Available at SSRN:; Zoltan Hajnal, Nazita Lajevardi, and Lindsay Nielson, “Voter Identification Laws and the Suppression of Minority Votes,” Journal of Politics 79 (April 2017): 363–379.

32. Stephen Ansolabehere, “Effects of Identification Requirements on Voting: Evidence From the Experiences of Voters on Election Day,” PS: Political Science and Politics 42, no. 1 (January 2009): 127–130; Lonna Rae Atkeson, Lisa Ann Bryant, Thad E. Hall, Kyle L. Saunders, and R. Michael Alvarez, “A New Barrier to Participation: Heterogeneous Application of Voter Identification Policies,” Electoral Studies 29 (March 2010): 66–73.

33. “State Voting Laws,” Brennan Center for Justice, accessed May 26, 2021,

34. Markus Prior, “You’ve Either Got It or You Don’t? The Stability of Political Interest Over the Life Cycle,” Journal of Politics 72 (July 2010): 747–766; Markus Prior, Post-broadcast Democracy: How Media Choice Increases Inequality in Political Involvement and Polarizes Elections (New York: Cambridge University Press, 2007); Michael X. Delli Carpini and Scott Keeter, What Americans Know About Politics and Why It Matters (New Haven, CT: Yale University Press, 1996); Larry M. Bartels, “Partisanship and Voting Behavior, 1952–1996,” American Journal of Political Science 44 (January 2000): 35–50.

35. Angus Campbell, Philip E. Converse, Warren E. Miller, and Donald E. Stokes, The American Voter (New York: John Wiley, 1960); Sidney Verba, Kay Lehman Schlozman, and Henry E. Brady, Voice and Equality (Cambridge, MA: Harvard University Press, 1995); Elizabeth Theiss-Morse, “Conceptualizations of Good Citizenship and Political Participation,” Political Behavior 15 (December 1993): 355–380.

36. Alan S. Gerber, Donald P. Green, and Christopher W. Larimer, “Social Pressure and Voter Turnout: Evidence From a Large-Scale Field Experiment,” American Political Science Review 102 (February 2008): 33–48.

37. Robert M. Bond, Christopher J. Fariss, Jason J. Jones, Adam D. I. Kramer, Cameron Marlow, Jaime E. Settle, and James H. Fowler, “A 61-Million-Person Experiment in Social Influence and Political Mobilization,” Nature 489: 295–298.

38. James H. Fowler, Laura A. Baker, and Christopher T. Dawes, “Genetic Variation in Political Participation,” American Political Science Review 102 (May 2008): 233–248; Christopher T. Dawes and James H. Fowler, “Partisanship, Voting, and the Dopamine D2 Receptor Gene,” Journal of Politics 71 (July 2009): 1157–1171; Robert Klemmensen, Peter K. Hatemi, Sara Binzer Hobolt, Inge Petersen, Axel Skytthe, and Asbjø;rn S. Nø;rgaard, “The Genetics of Political Participation, Civic Duty, and Political Efficacy Across Cultures: Denmark and the United States,” Journal of Theoretical Politics 24 (June 2012): 409–427; Robert Klemmensen, Peter K. Hatemi, Sara Binzer Hobolt, Inge Petersen, Axel Skytthe, and Asbjø;rn S. Nø;rgaard, “Heritability in Political Interest and Efficacy Across Cultures: Denmark and the United States,” Twin Research and Human Genetics 15, no. 1 (2012): 15–20.

39. The estimates of the percentages of those who never vote, the constant voters, and the transient voters come from Adam Berinsky, Nancy Burns, and Michael W. Traugott, “Who Votes by Mail? A Dynamic Model of the Individual-Level Consequences of Vote-by-Mail Systems,” Public Opinion Quarterly 65 (June 2001): 178–197. See also Berinsky, “Perverse Consequences of Electoral Reform in the United States.”

40. See, for example, Alan S. Gerber and Donald P. Green, “The Effects of Canvassing, Telephone Calls, and Direct Mail on Voter Turnout: A Field Experiment,” American Political Science Review 94 (September 2000): 353–363; Donald P. Green, Alan S. Gerber, and David W. Nickerson, “Getting Out the Vote in Local Elections: Results From Six Door-to-Door Canvassing Experiments,” Journal of Politics 65 (November 2003): 1083–1096; Alan S. Gerber, Donald P. Green, and Christopher W. Larimer, “An Experiment Testing the Relative Effectiveness of Encouraging Voter Participation by Inducing Feelings of Pride or Shame,” Political Behavior 32 (September 2010): 409–422; Kevin Arceneaux and David W. Nickerson, “Who Is Mobilized to Vote? A Re-analysis of 11 Field Experiments,” American Journal of Political Science 53 (January 2009): 1–16; David W. Nickerson, “Quality Is Job One: Professional and Volunteer Voter Mobilization Calls,” American Journal of Political Science 51 (April 2007): 269–282.

41. Robert B. Cialdini and M. R. Trost, “Social Influence: Social Norms, Conformity, and Compliance,” in The Handbook of Social Psychology, 4th ed., vol. 2, eds. D. T. Gilbert, Susan T. Fiske, and G. Lindzey (Boston: McGraw-Hill, 1998), 151–192; Robert B. Cialdini and Noah J. Goldstein, “Social Influence: Compliance and Conformity,” Annual Review of Psychology 55 (2004): 591–621.

42. Shang E. Ha and Dean S. Karlan, “Get-Out-the-Vote Phone Calls: Does Quality Matter?,” American Politics Research 37 (March 2009): 353–369; Nickerson, “Quality Is Job One.”

43. Charles Duhigg, “Campaigns Mine Personal Lives to Get Out Vote,” New York Times, October 14, 2012, A1.

44. “Messina and Zeleny Discuss 2021 Presidential Campaign,” Peter J. Hoagland Integrity in Public Service Lecture at the University of Nebraska–Lincoln, April 5, 2013, YouTube video posted April 10, 2013, accessed June 29, 2021,; see also Sasha Issenberg, “When It Comes to Targeting and Persuading Voters, the Obama Campaign Has a Massive, Insurmountable Advantage,” Slate, October 29, 2012,

45. Joshua Darr, “Where Clinton Is Setting Up Field Offices—and Where Trump Isn’t,” FiveThirtyEight (blog), October 7, 2016,; Susan Milligan, “The Fight on the Ground,” U.S. News and World Report, October 14, 2016,

46. Sean Sullivan, Michelle Ye Hee Lee, Anu Narayanswamy, and Josh Dawsey, “Trump Aims for Adulation. Biden Goes Virtual. The Two Presidential Candidates Are Running Vastly Different Campaigns as Election Day Nears,” Washington Post, October 21, 2020,

47. Arceneaux and Nickerson, “Who Is Mobilized to Vote?”

48. Larry M. Bartels and Lynn Vavreck, “Meet the Undecideds,” Campaign Stops (blog), New York Times, July 30, 2012,

49. Luciano Arcuri, Luigi Castelli, Silvia Galdi, Cristina Zogmaister, and Alessandro Amadori, “Predicting the Vote: Implicit Attitudes as Predictors of the Future Behavior of Decided and Undecided Voters,” Political Psychology 29 (June 2008): 369–387; but see Malte Friese, Colin Tucker Smith, Thomas Plischke, Matthias Bluemke, and Brian A. Nosek, “Do Implicit Attitudes Predict Actual Voting Behavior Particularly for Undecided Voters?,” PLoS ONE 7 (August 2012): e44130. doi:10.1371/journal.pone.0044130.

50. Nate Silver, “The Invisible Undecided Voter,” FiveThirtyEight (blog), January 23, 2017,

51. John R. Hibbing and Elizabeth Theiss-Morse, Stealth Democracy (New York: Cambridge University Press, 2002), 127.

52. E. J. Dionne Jr., “The Intensity Gap,” Washington Post, October 26, 2004, A25.

Descriptions of Images and Figures

Back to Figure

The horizontal axis ranges from 1860 to 2020 in increments of 4. The vertical axis is labeled percentage and ranges from 0 to 100 in increments of 10. The approximate data from the graph are tabulated below.

Back to Figure

The horizontal axis ranges from 1860 to 2020 in increments of 4. The vertical axis is labeled percentage and ranges from 0 to 100 in increments of 10. The line for congressional shows sharp fluctuations. The approximate data from the graph are tabulated below.

Back to Figure

The horizontal axis ranges from 1952 to 2020 in increments of 4. The vertical axis is labeled percentage and ranges from 0 to 100 in increments of 10. The approximate data from the graph are tabulated below.

Back to Figure

The horizontal axis ranges from 1952 to 2020 in increments of 4. The vertical axis is labeled percentage and ranges from 0 to 100 in increments of 10. The approximate data from the graph are tabulated below.

Back to Figure

The horizontal axis ranges from 1952 to 2020 in increments of 4. The vertical axis is labeled percentage and ranges from 0 to 100 in increments of 10. The approximate data from the graph are tabulated below.

Back to Figure

The horizontal axis ranges from 1952 to 2016 in increments of 4. The vertical axis is labeled percentage and ranges from 0 to 100 in increments of 10. The approximate data from the graph are tabulated below.

Back to Figure

The horizontal axis ranges from 1952 to 2020 in increments of 4. The vertical axis is labeled percentage and ranges from 0 to 60 in increments of 10. The approximate data from the graph are tabulated below.

If you find an error or have any questions, please email us at Thank you!