Michael W. Traugott

Web Survey Design and Administration

Many claims are being made about the advantages of conducting surveys on the Web. However, there has been little research on the effects of format or design on the levels of unit and item response or on data quality. In a study conducted at the University of Michigan, a number of experiments were added to a survey of the student population to assess the impact of design features on resulting data quality. A sample of 1,602 students was sent an e-mail invitation to participate in a Web survey on attitudes toward affirmative action. Three experiments on design approaches were added to the survey application. One experiment varied whether respondents were reminded of their progress through the instrument. In a second experiment, one version presented several related items on one screen, while the other version presented one question per screen. In a third experiment, for one series of questions a random half of the sample clicked radio buttons to indicate their answers, while the other half entered a numeric response in a box. This article discusses the overall implementation and outcome of the survey, and it describes the results of the imbedded design experiments.

REDUCING VOTE OVERREPORTING IN SURVEYS SOCIAL DESIRABILITY, MEMORY FAILURE, AND SOURCE MONITORING

One of the most frequently observed survey measurement errors is theoverreporting of voting behavior. Almost since the series of AmericanNational Election Studies (NES) began, the level of survey reported turn-out has been higher than estimates of turnout based on aggregate votetotals and census counts of the population (Clausen 1968). In addition,validation studies have confirmed that respondents will tend to overreportvoting (Abelson, Loftus, and Greenwald 1992; Parry and Crossley 1950;Presser, Traugott, and Traugott 1990; Silver, Anderson, and Abramson1986; Traugott and Katosh 1979).Two different explanations of overreporting have been tendered. Oneexplanation considers overreporting the result of social desirability, in

Who votes by mail? A dynamic model of the individual-level consequences of voting-by-mail systems

Election administrators and public officials often consider changes in electoral laws, hoping that these changes will increase voter turnout and make the electorate more reflective of the voting-age population. The most recent of these innovations is voting-by-mail (VBM), a procedure by which ballots are sent to an address for every registered voter. Over the last 2 decades, VBM has spread across the United States, unaccompanied by much empirical evaluation of its impact on either voter turnout or the stratification of the electorate. In this study, we fill this gap in our knowledge by assessing the impact of VBM in one state, Oregon. We carry out this assessment at the individual level, using data over a range of elections. We argue that VBM does increase voter turnout in the long run, primarily by making it easier for current voters to continue to participate, rather than by mobilizing nonvoters into the electorate. These effects, however, are not uniform across all groups in the electorate. Although VBM in Oregon does not exert any influence on the partisan composition of the electorate, VBM increases, rather than diminishes, the resource stratification of the electorate. Contrary to the expectations of many reformers, VBM advantages the resource-rich by keeping them in the electorate, and VBM does little to change the behavior of the resource-poor. In short, VBM increases turnout, but it does so without making the electorate more descriptively representative of the voting-age population.

Assessing the accuracy of polls and surveys.

Direct measurements of public opinion about national affairs appear with increasing frequency in all of the mass media. While such survey results are often with statements as to expected error margins, discrepancies between multiple surveys in the news at the same time on what seem to be the same topics may convince casual consumers that such error margins must be considerably understated. A brief review of the several sources of variability and fixed bias in such surveys provides a clearer frame of reference for the evaluation of such data.

The Impact of Voting by Mail on Voter Behavior

Most of the studies of voter behavior have dealt with voter turnout, but few have looked at other aspects of voting behavior that could be linked to balloting method. A reasonable amount of information has now accumulated about the impact of the shift from polling place elections to voting by mail on turnout, rolloff, drop-off, differences in voting for partisan offices and referenda, and differences in straight-ticket voting. This article analyzes recent time series of voting data in Oregon to assess the impact of the shift in voting method on these issues. The analysis includes data at the state, county, precinct, and individual levels, including individual ballots. The results suggest new criteria for evaluating shifts from one voting method to another that may be applied to other electoral reforms, such as those that will result from the Help America Vote Act.

THE POLLS—A REVIEW: EXIT POLLS IN THE 1989 VIRGINIA GUBERNATORIAL RACE: WHERE DID THEY GO WRONG?

In spite of common problems associated with sampling, question wording, and interviewing procedures, well-designed exit polls-in which voters are briefly interviewed as they leave the ballot booth-generally provide accurate projections of election outcomes (see, e.g., Mitofsky and Waksberg 1989). The fact that some recent exit polls have been considerably off the mark, then, has surprised many professionals. Moreover, recent misestimates of election outcomes have helped fuel public suspicion about the inaccuracies of polls in general (Clymer 1989; Davidson 1989). One very prominent incident occurred in the 1989 Virginia gubernatorial election. In that contest, Democrat L. Douglas Wilder narrowly defeated his Republican opponent, J. Marshall Coleman, winning by barely two-tenths of one percentage point. Yet an experienced research firm, Mason-Dixon Opinion Research (MDOR), conducted an exit poll for several television stations in Virginia and Washington, DC, and estimated an easy Wilder victory-by a 10-percentage-point margin (55-45 percent). This 5-percentage-point discrepancy was well outside the bounds of normal sampling error. In the same survey, however, MDOR accurately estimated an easy victory for the Democratic incumbent attorney general, Mary Sue Terry.

Early Appraisals of Electronic Voting

With the recent troubles in U.S. elections, there has been a nationwide push to update voting systems. States and localities are investing heavily in electronic voting systems, many of which use a touch screen. These systems offer the promise of faster and more accurate voting; however, the current reality is that they have some shortcomings in terms of voter usability. This study examines issues related to the usability of electronic voting systems and reports on a series of usability studies that involved expert review, close observation, a field test, and an exit poll to learn voters’responses to a new voting system. Our analysis shows these systems work well; however, they have some shortcomings including some that have raised concerns among a minority of voters.

THE IMPORTANCE OF PERSISTENCE IN RESPONDENT SELECTION FOR PREELECTION SURVEYS

Different techniques for respondent selection can af- fect data quality. These differences can result in variations in the distribution of partisans in preelection surveys, which in turn can have an effect on the distribution of candidate preference. Persis- tence in trying to interview designated respondents in telephone households increased the number of Republicans in a 1984 sam- ple, and therefore Reagans margin over Mondale. Such dif- ferences in interviewing techniques might account for some of the variations in national preelection estimates of the outcome of the presidential election, and they suggest that caution be used in comparing marginals for party identification from differ- ent surveys employing different respondent selection tech- niques. Appropriate sample designs are only a necessary prerequisite for col- lecting valid data with the survey method. The implementation of the sample design also has an important effect on the quality of the data which are obtained. Variations in techniques for selecting respondents within sample households can affect the quality of data and therefore the substantive conclusions which are drawn from many surveys. These effects derive from a number of factors, including the way in which individuals residing in the household are enumerated or listed, if at all, how a designated respondent is subsampled, and the persistence with which designated respondents are called back in order to obtain an interview. While these methodological differences affect the demo- graphic characteristics of the sample, the most significant conse- MICHAEL W. TRAUGOTT iS Senior Study Director, Center for Political Studies, The Uni- versity of Michigan. An earlier version of this article was presented at the Annual Meeting of the American Association for Public Opinion Research, McAfee, New Jersey, May 16-19, 1985. The research assistance of Julio Borquez is gratefully acknowl- edged.

Mass Media in U. S. Congressional Elections

Congressional candidates in the United States run campaigns that are largely independent of national parties and their leaders. The U.S. media are free of party control and decentralized geographically. Many legislative candidates buy time on local radio and television to air campaign ads, and they seek news coverage as well. Incumbents have inherent advantages in these activities, although these are greater in House than in Senate elections. As a result, congressional candidates tend to become relatively visible over the course of the campaign, depending upon the degree of overlap of their constituency boundaries and media markets as well as upon the competitiveness of their races. What U.S. voters learn about candidates for legislative office depends substantially on where the voters live.

Losing fewer votes: The impact of changing voting systems on residual votes

Problems in the 2000 presidential election, especially in Florida, initiated a large-scale shift toward new voting technology. Using cross-sectional and longitudinal data, we report on the effects of changes in voting systems in Florida and Michigan. The variety of initial conditions and the numerous changes make these excellent case studies. We find that reforms succeeded in reducing the residual vote. Every change from old to new technology resulted in a decline in residual votes that was significantly greater than in areas that did not change voting equipment. The percentage of residual votes in the 2004 presidential race in localities that changed voting systems was well under 1 percent, representing a 90 percent reduction in error in Florida and a 35 percent reduction in Michigan. We run these analyses separately for undervotes and overvotes. Using ecological-inference techniques, we investigate the persistence of residual votes when technology changed and find very little persistence.