Analyzing Voting Behavior

Charles Prysby and Carmine Scavo, American Voting Behavior in Presidential Elections: 1972 to 1992 (Washington, D.C.: American Political Science Association, 1993).

Analyzing the behavior of voters in recent presidential elections requires a theoretical understanding of the basic forces that influence electoral behavior. Two major concerns characterize the research in this area. One concern is with explaining the results of particular elections by identifying the sources of individual voting behavior. We attempt to understand an election outcome by analyzing how and why the voters made up their minds. A second major concern in voting research focuses on the dynamics of electoral behavior. In this case, the emphasis is on the nature and sources of changes in voting patterns over time. These two concerns are complementary, not contradictory, but they do emphasize different sets of theoretical questions. For our purposes, these two concerns provide a useful basis for discussing key aspects of voting behavior.

Sources of Individual Voting Behavior

On what basis do voters decide how they will cast their ballot? Several basic factors can be identified as reasons for preferring one candidate to the others in an election. The voter may decide on the basis of one or more of the following considerations: (a) evaluations of the personal qualities of the candidates; (b) general assessments of the performance of government; (c) orientations on specific issues of public policy; and (d) basic loyalty to or preference for a particular political party. Candidate qualities and government performance are distinctly short-run forces, capable of substantial shifts from one election to the next. Party loyalties are much more stable in the short run. Issue orientations fall somewhere in between. While the specific issues crucial in presidential elections can change dramatically, many basic policy questions (e.g., defense spending, welfare programs, abortion) stretch across several elections, with partisan differences on these issues remaining relatively constant.

Voters frequently judge presidential candidates by their personal characteristics. They form images of the personal qualities and abilities of the candidates, and these perceptions are important influences on the vote. Included among the relevant characteristics are such things as the knowledge, experience, honesty, morality, compassion, competence, and leadership ability of the candidates (Page 1978). Such considerations seem to have played an important role in many recent presidential campaigns. In 1992, Bill Clinton was accused of marital infidelity and of avoiding the military draft during the Vietnam War, while President Bush was attacked for being a weak leader. Similarly, Michael Dukakis was criticized by Bush in 1988 for being unpatriotic and lacking American values, while questions were raised about Bush's leadershin qualities by the Democrats. While some candidates have been hurt by negative perceptions of their personal characteristics, others have benefited from very positive ones. For example, much has been made of Ronald Reagan's strong personal appeal to voters.

There is substantial variation from one election to the next in voter perceptions of the personal characteristics of the presidential candidates. The Republican candidate may have a significant advantage when it comes to personal characteristics in one year, while the opposite may be true in the following election. The uncertainty in knowing very far ahead of the election how the presidential candidates will be perceived in personal terms is one factor that makes predicting the outcome of presidential elections a difficult task. A number of items in the dataset for this module deal with respondent evaluations of the personal qualities of the presidential candidates, allowing us to examine the influence of these factors in particular elections and to compare this influence across elections.

The role of public policy issues in elections is of particular interest to political analysts. The term issue sometimes is used loosely to refer to anything that is a source of conflict or contention, but that is not its meaning here. We are referring to public policy issues, meaning questions of what the government should or should not do. Policy issues involve conflict over the direction of government action. Some policy issues in an election may be quite specific, such as the conditions under which abortion should be legal. Other policy issues are general, dealing with broad approaches to problems, such as whether the federal government should cut spending on social welfare programs. Elections are widely justified as providing a means for citizens to influence governmental decisions by choosing among contenders for office, and the assumption often is that the electorate will shape government policy by selecting candidates on the basis of policy issues. When this does not appear to be the case, political commentators often are quite critical, and we frequently hear complaints that the candidates in an election are failing to clearly address the issues.

For a policy issue to affect the vote decision, voters must have opinions on the issue and must perceive differences between the candidates on the issue. Even on important issues, some voters will fail to meet these conditions. Some will have opinions that are too weak and unstable to provide a basis for evaluating the candidates, while others will fail to see any significant differences between the candidates on the issue (Campbell et al. 1960, 167-187). But many voters will have definite opinions and clear perceptions of candidate differences on at least some issues, and this will be true especially when the candidates clearly articulate their differences (Nie, Verba, and Petrocik 1976, 156-173).

Specific issues have received considerable discussion in many recent presidential campaigns. Health care, for example was a key issue in 1992, as Clinton called for a national health care plan. Defense spending has been important in some campaigns, especially 1980, when Reagan argued that substantial increases were needed in defense spending to restore American military strength. Abortion has divided the presidential candidates in a number of recent elections. Taxes and government spending almost always receive considerable attention in presidential campaigns. The data for this module contain measures of how respondents felt about a number of very basic policy issues, allowing us to examine the impact of these attitudes on the vote in recent presidential elections.

Although the specific policy issues of greatest importance vary considerably from one election to the next, the general pattern has been for the Democratic candidate to be perceived as more liberal overall and the Republican candidate more conservative overall. The extent of the differences between the candidates has varied. In 1976, the two candidates were seen in fairly moderate terms, with the Democrat slightly left of center and the Republican slightly right of center (Abramson, Aldrich, and Rohde 1982, 126129). In other years, such as 1984, the ideological differences were much greater (Abramson, Aldrich, and Rohde 1987, 168-172).

Rather hall choosing among candidates on the basis of specific issues of public policy, voters may rely more on general evaluations of the performance of government. A presidential election is, at least in part, a referendum on the performance of the incumbent administration. This is especially true when the incumbent president is running for reelection, which was the case in all but one of the elections held between 1972 and 1992. In the one exception, 1988, the incumbent Vice President was the presidential candidate of his party. Retrospective evaluations of government performance are an important determinant of voting behavior, and this effect should be distinguished from the influence of policy issues (Fiorina 1981). Policy issues involve differences over what the government should do; they are prospective in nature. Performance evaluations involve differences over how well the government has done; they are retrospective in nature. Quite often, we find agreement over what the government should accomplish, but disagreement over how well the goals have been achieved. Basic goals such as low unemployment, low inflation, steady economic growth, national security, and world peace are shared by all. Candidates do differ in their prescriptions for economic health or national security, but discussions of the details of macroeconomic theory, or of defense strategies, may not be scrutinized by many voters. More relevant are general perceptions of whether the economy or national security has improved or declined.

The importance of these factors is reflected by the emphasis given to them in recent presidential elections. The condition of the economy was a critical factor in the 1992 presidential election. Bush's opponents claimed that he was responsible for the poor health of the economy, while the President claimed that the nation's economic problems were being exaggerated by the media. Republicans in 1980 sought to tie negative evaluations of the economy and the international environment to perceptions of President Carter's competence. In 1984, the Republicans emphasized the improvement in the economy and the international environment that occurred during the Reagan administration. Similar claims were made by Republicans in 1988, while Democrats countered that everything was not so well off.

Although evaluations of government performance can involve more than just the condition of the economy, it appears that economic conditions are particularly important in influencing the performance evaluations of voters (Kiewiet 1983). Moreover, economic conditions tend to shift dramatically. In some years, such as 1980 or 1992, the economy was badly off, especially in the eyes of the public, whereas in other years, such as 1972 or 1984, there was considerable satisfaction with the state of the economy. Still other years, such as 1976 or 1988 received more mixed evaluations from the public. This dataset includes several measures of perceptions of government performance and related factors, including items dealing with economic conditions, thereby allowing for an examination of the role of retrospective evaluations in presidential elections.

Finally, a voter may cast a ballot for a candidate because of the candidate's party affiliation. Most voters express some identification with one of the two major political parties, although these attachments tend to be substantially weaker than in the past. These basic partisan loyalties may influence their votes. Party identification normally is measured by asking individuals whether they consider themselves to be a Democrat, Republican, or independent. Those indicating Democratic or Republican can then be asked whether they are a strong or a weak Democrat or Republican, while those claiming to be an independent can be asked whether they feel closer to one of the two political parties. This yields a sevenfold classification: strong Democrat, weak Democrat, independents closer totheDemocrats, independents not closer to either party, independents closer to the Republicans, weak Republicans, and strong Republicans.

The direct influence of party identification on the vote probably is small in presidential elections. By this we mean that there probably were few voters in recent elections who cast their ballots for or against a presidential candidate simply because he was a Democrat or a Republican. But the indirect influence of party identification is much greater, in that partisan loyalties influence evaluations of candidates, assessments of government performance, and perceptions of political events. Put simply, party identification is a perceptual screen-a pair of partisan tinted eyeglasses through which the voter views the political world. Party identification is less important now than in the past, but it is still a significant factor for explaining political orientations and behavior. In fact, party identification is still a very good predictor of how individuals vote.

Evaluations of candidate personal characteristics, assessments of government performance, and orientations on public policy issues are not only influenced by party identification. They affect each other. Strong conservatives are not only more likely to agree with the issue positions of a conservative candidate; they also will be likely to have a favorable view of the personal characteristics of that candidate. Voters who feel that the incumbent President has not managed the economy very well are likely to unfavorably judge many of the President's personal qualities, especially those regarding leadership. Although the three attitudes discussed above are conceptually distinct, they are empirically interrelated. In sum, there are multiple reasons why people vote the way that they do. The relative weight of these factors may vary among individuals. Some people may vote more on the basis of the personal characteristics of the candidates, others more on the basis of issues of public policy, and still others on the basis of retrospective performance evaluations. A few may cast a ballot largely on the basis of party identification. The importance of the above factors also can vary from one election to another. While it is easy to outline the possible factors that motivate the vote, it is more difficult to specify exactly how important each factor was in determining the election outcome.

Some of the important controversies in the literature on voting behavior involve the relative importance of different factors in determining individual voting. During the 1950s, considerable emphasis was placed on the role of party identification as a determining factor (Campbell et al. 1960). Voters were viewed as politically unsophisticated, responding primarily to basic partisan loyalties and lacking in an understanding of the issues underlying the election. During the 1960s and 1970s, the literature began to focus more on the role of issues in elections (Key 1966; Nie, Verba, and Petrocik 1976). Voters were now seen as capable of choosing among candidates on the basis of issues. In the 1980s, considerable attention was paid to retrospective voting, especially in response to economic conditions (Fiorina 1981; Kiewiet 1983). To some extent, these changes in emphasis reflected changes in voting behavior, but they also represented changes in the theories and methodologies used to study voting behavior.

Controversies in explaining voting behavior have encompassed one of the most basic concepts in the field, party identification. Earlier literature stressed that party identification was learned at a relatively early age and remained relatively stable throughout the life span (Campbell et al., 1960). Once formed, party identification exercised considerable effect on an individual's political attitudes, but there was little influence in the opposite direction. Recently there has been more acceptance of an alternative view of party identification. This conception sees party identification as reflecting as well as influencing individual attitudes. Voters can and do shift their party identifications in response to their evaluations of the parties in terms of issues and performance. Even this revised view conceives of party identification as a fairly stable orientation-more stable than issue orientations or performance appraisals, for example but it gives more consideration to how individuals might bring their party identification into line with other attitudes.

Analyzing the factors that affect the vote decision also leads to a consideration of how voters form their perceptions and evaluations. Why, for example, do some voters feel that a candidate is trustworthy, while others have the opposite opinion? What leads voters to have favorable or unfavorable assessments of government economic performance? As suggested above, party identification plays a role, but it does not fully explain the formation of these attitudes and orientations. Members of the same party often will have divergent views. While we do not have the space here for an extended discussion of this topic, it is clear that personal experiences, basic values, membership in various groups, exposure to the mass media, and political discussions with friends all play an influential role. One set of factors, social and demographic characteristics, are frequently examined, in part because they form basic cleavages in society. Race, religion, region, and socio-economic status have divided voters for decades. Other characteristics, such as gender or age, have played a role in recent elections as well. The influence of these social and demographic characteristics on attitudes and voting is due to the fact that they shape the experiences, values, interests, and information sources to which the voters are exposed.

Electoral Dynamics

Election results often change dramatically. A lopsided victory for one party may be followed by a landslide for the other party in the following election, as the discussion in Chapter I shows. Electoral changes can be divided into two types: short-run and long-term. Short-run changes are the result of fluctuations in factors that are specific to an election, such as the characteristics of the candidates or the particular issues that are salient. These short-term factors may be moderately favorable to the Democrats in one election, strongly favorable to the Republicans in another, and evenly divided in a third.

Voter perceptions of candidate personal characteristics and evaluations of government performance are primarily short-run factors. They tend to be specific to a particular election, and they easily change from one election to the next, as the candidates and events change. The fact that one party has a big advantage on these two factors in one election does not mean that it will have a similar advantage in the following election. However, these perceptions and evaluations can have long-term effects. Voters may develop an identification with a party because of favorable attitudes toward the candidates, policies, and accomplishments of the party-or because of unfavorable attitudes toward the opposing party. A voter's party identification may represent, at least in part, a summary evaluation of how the voter has perceived recent political history (Fiorina 1981).

Long-term shifts result from alterations in basic loyalties and represent changes that last beyond a particular election. The most significant long-term change occurs when there is a critical realignment of the party system, which refers to a relatively rapid, fundamental, and durable alteration in the pattern of party loyalties (Sundquist 1983). Realignments occur infrequently; the last major upheaval of the party system occurred in the 1930s, and before that in the 1890s and 1850s. Of course, in any time period there is some change in party loyalties, but only rarely is it substantial enough to qualify as a realignment.

The 1930s realignment shaped the current party system. The Great Depression acted as the catalyst for a transformation of the party system that moved the Democrats from minority to majority status at the national level. The New Deal Democratic coalition that put Franklin D. Roosevelt in the White House and the Democratic Party in control of Congress combined support form the working class and various ethnic and minority groups with already existing strength in the South. The basis of Democratic appeal to blue-collar workers, low-income individuals, and recent immigrant groups (largely Catholics and Jews from southern and eastern Europe) was the party's liberalism in economic matters. Roosevelt and the Democrats favored federal government activity to combat the Depression and proposed programs to benefit disadvantaged groups. The Republicans, who appealed more to the middle-class, business groups, and northern white Protestants, were critical of this expansion of government interference in the economy and creation of a variety of social welfare programs. By the late 1930s, the lines between the two parties were clearly drawn, both in ideological and socioeconomic terms (Ladd and Hadley 1978).

Although the impact of the New Deal realignment has remained for decades, it has done so only in a diluted and revised form. The class cleavages that were so clear in the 1930s and 1940s have diminished greatly in recent years (Abramson, Aldrich, and Rohde 1995, 146). Differences between Protestants and Catholics in their partisanship have eroded somewhat as well (Abramson, Aldrich, and Rohde 1995, 146). And the South, once solidly Democratic, has voted disproportionately Republican in recent elections (Black and Black 1992, 58-76). Still, the party images of an earlier era persist. Democrats remain thought of as the party that favors bigger government, more spending on domestic programs, and helping those at the bottom of the economic pyramid. Republicans continue to be perceived as favoring limited government, less spending on domestic programs, and fewer restrictions on business enterprises. These are not baseless images. They reflect continuing fundamental differences between the parties.

At the same time that many of the old partisan differences have diminished, new divisions have emerged. Beginning in the 1960s, blacks began to vote in greater numbers and to cast their ballots overwhelmingly for Democratic candidates. They now are one of the most loyal components of the Democratic coalition (Abramson, Aldrich, and Rohde 1995, 144-147). This development may be one factor in the declining support than Democrats have received from white voters (Edsall and Edsall 1991). In the 1980s, a gender gap emerged, with men more likely than women to vote Republican (Mueller 1988). Marital status also has become important in recent elections; married individuals now are more Republican in their voting than are single people (Weisberg 1987). A new religious split has been developing, with Republicans appealing more to a "New Christian Right," which emphasizes traditional moral values (Wald 1987). And much has been made of Republican tendencies among suburban voters, who constituted a majority of the electorate for the first time in 1992 (Schneider 1992).

While there has been a great deal of change in recent years, it has not added up to a fundamental realignment, on the order of what happened in the 1930s. The changes have occurred more slowly and have been less complete. In fact, some of what has occurred might be better described as dealignment: many voters have become less attached to either party, with the result being a very volatile electorate that is highly responsive to short-term forces (Wattenberg 1994). The diminished loyalty to parties is reflected in increased ticket splitting (voting for Democrats for some offices and Republicans for others). It also may be reflected in a greater willingness to vote for a third-party or independent candidate, such as Ross Perot. Still, over the 1970s and 1980s the Democratic advantage in partisan loyalties eroded considerably, to the point that Republicans were close to parity among those who actually voted (Asher 1992, 86).


A number of attitudinal and social factors are related to individual voting behavior. Among attitudinal factors, assessments of the personal characteristics of the candidates, evaluations of the performance of the government, orientations on specific policy issues, and party identification are the primary determinants of candidate choice. For social factors, race, religion, region, social class, and gender appear to be the characteristics most closely related to voting. Examining how these factors are related to the vote in particular elections not only allows us to explain the election outcome, but also can provide us with an understanding of electoral dynamics. All of the ideas raised in this chapter can be examined, at least to a certain extent, with the data contained in this package.


Abramson, Paul R., John H. Aldrich, and David W. Rohde. 1982. Change and Continuity in the 1980 Elections. Washington, DC: Congressional Quarterly Press.
Abramson, Paul R., John H. Aldrich, and David W. Rohde. 1987. Change and Continuity in the 1984 Elections. Revised edition. Washington, DC: Congressional Quarterly Press.
Abramson, Paul R., John H. Aldrich, and David W. Rohde. 1991. Change and Continuity in the 1988 Elections. Revised edition. Washington, DC: Congressional Quarterly Press.
Abramson, Paul R., John H. Aldrich, and David W. Rohde. 1995. Change and Continuity in the 1992 Elections. Revised edition. Washington, DC: Congressional Quarterly Press.
Asher, Herbert B. 1992. Presidential Elections and American Politics. Pacific Grove, CA: Brooks/Cole Publishing Co.
Black, Earl, and Merle Black. 1992. The Vital South. Cambridge: Harvard University Press.
Campbell, Angus, Philip E. Converse, Warren E. Miller, and Donald R. Stokes. 1960. The American Voter. New York: John Wiley and Sons.
Converse, Philip. 1964. "The Nature of Belief Systems in Mass Publics," in Ideology and Discontent, edited by David E. Apter. New York: The Free Press, 1964.
Converse, Philip. 1970. "Attitudes and Nonattitudes: The Continuation of a Dialogue," in The Quantitative Analysis of Social Problems, edited by Edward Tufte. Reading, MA: AddisonWesley.
Converse, Philip and Gregory Markus. 1979. "la Plus Ca Change ... : The New CPS Election Study Panel," American Political Science Review 73 (March), 32-49.
Edsall, Thomas B., and Mary D. Edsall. 1991. Change Reaction. New York: W.W. Norton.
Feigert, Frank B. 1993. "The Ross Perot Candidancy and Its Significance." America's Choice: The Election of 1992. William Crotty, editor. Guilford, CT: Dushkin.
Fiorina, Morris P. 1981. Retrospective Voting in American National Elections. New Haven, CT: Yale University Press.
Groves, Robert. 1989. "Actors and Questions in Telephone and Personal Interview Surveys," in Survey Research Methods: A Reader, edited by Eleanor Singer and Stanley Presser. Chicago: University of Chicago Press.
Key, V. O. , Jr. 1949. Southern Politics in State and Nation. New York: Alfred A. Knopf.
Key, V. O., Jr. 1966 The Responsible Electorate. Cambridge: Harvard University Press.
Kiewiet, D. Roderick. 1983. Macroeconomics and Micropolitics. Chicago: University of Chicago Press.
Kish, Leslie. 1949. "A Procedure for Objective Respondent Selection within the Household," Journal of the American Statistical Association, 44, pp. 380-387.
Ladd, Everett Carll Jr., and Charles D. Hadley. 1978. Transformations of the American Party System. Revised edition. New York: W.W. Norton.
Lamis, Alexander P. The Two-Party South. New York: Oxford University Press.
Lewis-Beck, Michael S., and Tom W. Rice. 1992. Forecasting Elections. Washington, DC: Congressional Quarterly Press.
Luttbeg, Norman R., and Michael M. Gant. 1995. American Electoral Behavior, 1952-1992. Itasca, IL: F.E. Peacock Publishers.
Miller, Arthur H., Warren E. Miller, Alden S. Raine, and Thad A. Brown. 1976. "A Majority Party in Disarray: Policy Polarization in the 1972 Election." American Political Science Review 70:753-778.
Mueller, Carol M. 1988. The Politics of the Gender Gap. Newbury Park, CA: Sage Publications.
Nie, Norman H., Sidney Verba, and John R. Petrocik. 1976. The Changing American Voter. Cambridge: Harvard University Press.
Page, Benjamin I. 1978. Choices and Echoes in Presidential Elections. Chicago: University of Chicago Press.
Pomper, Gerald M., et al. 1977. The Election of 1976. New York: David McKay.
Pomper, Gerald M., et al. 1981. The Election of 1980. Chatham, NJ: Chatham House.
Pomper, Gerald M., et al. 1985. The Election of 1984. Chatham, NJ: Chatham House.
Pomper, Gerald M., et al. 1989. The Election of 1988. Chatham, NJ: Chatham House.
Pomper, Gerald M., et al. 1993. The Election of 1992. Chatham, NJ: Chatham House.
Quirk, Paul J., and Jon K. Dalager. 1993. "The Election: A `New Democrat' and a New Kind of Presidential Campaign." The Elections of 1992. Michael Nelson, editor. Washington, DC: Congressional Quarterly Press.
Rogers, Theresa. 1989. "Interviews by Telephone and in Person: Quality of Response and Field Performance," in Survey Research Methods: A Reader, edited by Eleanor Singer and Stanley Presser. Chicago: University of Chicago Press.
Schneider, William. 1992. "The Suburban Century Begins." The Atlantic (July), 33-44.
Sundquist, James. 1983. The Dynamics of the Party System. Washington, DC: The Brookings Institution.
Wald, Kenneth D. 1987. Religion and Politics in the United States. New York: St. Martin's Press.
Wattenberg, Martin. 1994. The Decline of American Political Parties 1952-1992. Cambridge: Harvard University Press.
Weisberg, Herbert. 1987. "The Demographics of a New Voting Gap: Marital Differences in American Voting." Public Opinion Quarterly 51:335-343.
Weisberg, Herbert, Jon Krosnick, and Bruce Brwen. 1989. An Introduction to Survey Research and Data Analysis, 2nd edition. Glenview, IL: Scott, Foresman, and Company.
White, John K. 1993. "The General Election Campaign: Issues and Themes." America's Choice: The Election of 1992. William Crotty, editor. Guilford, CT: Dushkin.