'Who Votes?' Notes

1. Party identification:

A citizen's party identification may be thought of as an enduring tendency to support a slate of candidates from one's own party. Distinct from party registration or actual work on behalf of a party, this attitudes stems from a variety of sources.

Many scholars find that party identification starts to develop during early childhood, when psychological orientations towards the parties are passed down from parents to children. Key political crises -- such as the Great Depression of the 1930s or the protest movements of the 1960s -- may also shape the citizen's attitudes towards Democrats and Republicans. Furthermore, an individual's party identification undoubtedly reflects a commitment to some of the "issues of the day." Democratic identifiers, for example, might support some of the policies advanced by Democratic leaders in government, like affirmative action programs, funding to help the unemployed, or limits on spending for national defense.

In general, citizens who feel a strong attachment to the Democratic or Republican Party are more inclined to participate during elections. This is evident if we consider reported turnout in 1992 and 1988:

Political ideologies are sets of beliefs and values pertaining to how governments operate, how they should operate, and what kinds of policies they should produce. In the U.S., citizens who think in ideological terms usually identify themselves as being either conservative or liberal.

Ideological labels mean different things to different people. For example, conservatives might espouse any or all of the following views: a belief that government should not extend welfare benefits to poor people or attempt to redistribute income among citizens; hesitance to break with past traditions and moral codes; opposition to affirmative action programs favoring racial minorities in employment and higher education; or a willingness to use force internationally -- rather than negotiations through international organizations -- to protect the strategic or economic interests of the United States.

Often it is the case that ideological extremism covaries with voter turnout. Note, for example, the following distributions from 1992 and 1988:

Typically, 15 to 30 percent of the American electorate expresses great interest in politics. This figure is comparable to levels found in other advanced industrial democracies like Great Britain, France, and Germany.

Not surprisingly, citizens who are interested in politics are more inclined to discuss national issues with family and friends, contribute to one of the major parties, express strong ideological preferences, and, of course, vote on election day. This latter relationship is depicted neatly below using information collected during the 1992 and 1988 presidential campaigns:

All things being equal, older Americans tend to vote more often than younger Americans. Several factors might cause this distinction. Younger people tend to change addresses more often than older, more established citizens. This means that it is somewhat harder for young individuals to register to vote. Moreover, political issues, parties, and candidates are often not especially important for younger people. When citizens pass from their 20s into their 30s and 40s, however, they take on new social and economic roles. As they get situated in families, careers, and other social networks, people may grow more attentive to government policies. Research has shown that party affiliations grow stronger with age, as does ideological identification. Voting provides such citizens with an opportunity to act on these beliefs.

To illustrate this dynamic, consider the following table:

Education level is one of the most significant determinants of electoral participation, as the table below indicates. Citizens with a college degree are far more likely than the less educated to vote. Those with advanced education often know more about how the government works. They read newspapers more often, they have at their disposal more factual information about issues and candidates, and they are more aware of the consequences of government policy for their own lives. Furthermore, educated people more readily accept the norms and values of democratic life. For instance, they are usually more inclined to believe that "your vote counts," and that a citizen who does not participate has no moral right to criticize governmental actions.

Prior to federal legislation in the 1960s, race was a profound factor affecting turnout, especially in the southern United States. Literacy tests, grandfather clauses, and other kinds of legal barriers had effectively kept large numbers of African-Americans from participating. As recently as 1952, only 4 percent of southern blacks of voting age voted. Today, no such legal barriers remain. Nevertheless, we find that racial minority group members still report voting less often than white Americans, as the results below indicate. When socio-economic status is statistically controlled, however, the unique effect of race becomes considerably less than this table suggests.

In general, the more affluent the citizen, the greater the chance that he or she participates in elections. Several forces might cause this effect. Clearly, affluent people tend to have more free time to devote to political causes. Such individuals are also more likely to take an interest in politics, to believe that their policy preferences will be heard by policy makers, and to interact with others who also participate in greater numbers than poorer citizens.

By way of illustration, consider the following breakdown:

People residing in the southern part of the United States are less inclined than other Americans to vote, as the distributions below indicate. Abstentions in this part of the country may be due to several factors. Historically, local leaders prevented blacks from voting through the use of grandfather clauses, literacy tests, poll taxes, and open intimidation. Such measures -- which are all illegal now -- also worked to depress the turnout rates of poor whites. Additionally, lower voter participation in the south might be due to the traditional dominance of the Democratic Party in that region. All things being equal, turnout rates rise as competition between parties increases.

It is also worth noting that southerners on average are less affluent and less educated than citizens from other regions (and, of course, education and income significantly influence voter turnout). In one survey from 1992, for example, 23 percent of southern respondents reported having family incomes above $45,000, and 21 percent stated that they had received a college degree. For non-southerners, 33 percent reported family income levels above $45,000, and 25 percent were college educated.

Historically speaking, women were given the right to vote relatively recently, via the 19th Amendment to the U.S. Constitution (1920). Surveys conducted since 1952 show that women report voting less than men (see below). This disparity may be due to a number of factors. Up until recently, women on average were not as well educated as men, and education is clearly a predictor of voter turnout. Until the advent of "women's issues" and an increase in the number of women in the workplace, women did not pay as much attention to politics. Furthermore, women have tended to lack confidence in their ability to influence political decisions.

In spite of these traditional impediments to participation, however, we find today that women are slightly more likely to report voting once other demographic factors are statistically controlled. Thus, electoral biases against women as a group have largely fallen by the wayside.

Numerous surveys have shown Americans to be more religious on average than citizens of most other industrial democracies. In a survey conducted 1992, two-thirds of the respondents reported attending religious services. In general, people who attend churches or synagogues are more inclined to participate in electoral politics (see below). Why do we find such a relationship? Some researchers cite a kind of "spillover" effect at work: participation in social groups leads to participation in politics. As individuals interact within religious organizations, they often exchange information about political issues and candidates. They also might develop a sense of common purpose, which ultimately could then lead to increased political mobilization.

The connection between black churches and the civil rights movement in the 1950s and 1960s clearly illustrates this dynamic. Similarly, and more recently, many churches in the United States have encouraged their members to become politically active to support "family" issues -- e.g., prayer in public schools, an end to legalized abortion, or the continued exclusion of homosexuals from mainstream society.

Alexis de Tocqueville, the well-known social commentator from France, observed in the 1830s that Americans were remarkably willing to join civic organizations for the promotion of particular issues or causes. Though the United States has changed considerably since Tocqueville's writings, citizens continue to participate in community groups. In one survey from 1992, for example, 23 percent reported working with an organization during the previous twelve months.

Not surprisingly, participation in such groups correlates quite strongly with turnout rates (see below). When individuals meet to pursue common goals, they exchange political informationand resources. They also become more able to see the consequences of their activities, and they undoubtedly follow news reports about political issues and campaigns more closely.

Many states today have registration laws which require that prospective voters reside within a particular district for a minimum number of days, typically 30 or 60 days. This being the case, newcomers to an area are less likely to be eligible to vote. Moreover, citizens without a deep attachment to a community are less likely to become informed about local issues and candidates, and consequently their turnout will be lower than more established residents. Such a relationship is clearly evident in the table below:

The term "civic duty" may be thought of as a political responsibility citizens feel they have based on some emotional or symbolic tie to the larger community. Aside from partisanship, ideological orientation, demographic characteristics, and participation in social organizations, a sense of civic duty undoubtedly spurs many people to participate in an election. The item asking whether one would be happy to serve on a jury is a "marker" for this sense of civic duty. Persons willing to take time out of their day to serve on a jury (61 percent of the sample below) were also considerably more likely to participate in the 1992 election:

Feeling that one has some influence over what policy makers do is a central attitudinal linchpin of democracy. Political scientists refer to this belief as "external political efficacy." Unfortunately, in recent years only about half of the respondents to public opinion surveys have disagreed with the statement, "People like me don't have much say about what the government does."

Clearly, this item is related to turnout during elections. Citizens believing that they have no political influence are much less likely to vote:

For democracy to function properly, the citizen must have access to information about candidates, parties, and key public policy issues. Citizens must also believe that they have an adequate understanding of the issues facing the country, or at least that they have the potential to gain such an understanding. Political scientists label this subjective sense of confidence "internal political efficacy."

Numerous studies have shown that those who doubt their ability to make sense out of politics are less likely to vote. Consider, for example, the following distributions from the 1992 and 1988 campaigns:

The information used to develop this simulation comes from responses to the 1992 American National Election Study, conducted by the Center for Political Studies at the University of Michigan and made available through the Inter-university Consortium for Political and Social Research. [Neither the Center for Political Studies nor the ICPSR are responsible in any way for the analysis conducted in this simulation.]

Each response in this simulation is weighted in accordance with coefficients estimated via a logistic regression analysis for the National Election Study sample (N = 1,972). The dependent variable in this regression was self-reported voting participation. The full regression model appears below, with standard errors in parentheses. In total, 82.9 percent of the cases were correctly classified by the model, which represents a 28.9 percent proportional reduction in error.