Revisiting public opinion on voter identification and voter fraud in an era of increasing partisan polarization.

• Author: Stewart, Charles, III

Introduction I. The Changing Partisan Structure of Support for Voter Identification Laws II. Beliefs About the Prevalence of Voter Fraud A. Beliefs About the Prevalence of Voter Impersonation B. Perceptions of Fraud and the Likelihood of Voting C. Perceptions About Fraud and Voter Confidence III. Knowledge of Voter Identification Laws and Beliefs About Their Effects Conclusion Appendix A Appendix B Appendix C Introduction

Eight years ago, we published the first study on the relationship between voter identification laws and public attitudes concerning voter fraud. (1) That article, like this one, was motivated by the unique constitutional argument gaining favor in the courts positing that voter identification laws could be justified as a measure to instill voter confidence in elections. (2) Some advocates and judges argued that such laws would do so by convincing voters that, whatever the reality, such laws decrease voter fraud at the polls. We demonstrated then that public perception of voter fraud was unrelated to the presence or absence (or stringency) of voter identification laws. (3) Although voter identification laws were popular among the general population, as well as among subsets based on race and party, such laws did not have any appreciable impact on voters' perceptions of the prevalence of voter fraud or on voter turnout. (4) We revisit those findings here with the benefit of new public opinion data and more experience with voter identification laws in a greater number of states.

One additional motivating factor behind this Article is the debate concerning our earlier article that occurred among judges in the Seventh Circuit Court of Appeals in a recent challenge to Wisconsin's voter identification law, Frank v. Walker, (5) Striking down the law, the district court in that case had credited expert testimony that relied on our earlier article. (6) The panel opinion reversed, dismissing our Harvard Law Review study as not being published in a "refereed scholarly journal" and as "report[ing] the results of one opinion poll of people living throughout the country." (7) According to the panel, "[i]f the public thinks that photo ID makes elections cleaner, then people are more likely to vote or, if they stay home, to place more confidence in the outcomes. These are substantial benefits." (8) In a dissent joined by four other judges from the denial to rehear the case en banc, Judge Posner relied on our article to argue:

[T]hese laws do not reduce such fraud, for if they did one would expect perceptions of its prevalence to change. The study also undermines the suggestion in the panel's opinion (offered without supporting evidence) that requiring a photo ID in order to be allowed to vote increases voters' confidence in the honesty of the election, and thus increases turnout. If perceptions of the prevalence of voter-impersonation fraud are unaffected by the strictness of a state's photo ID laws, neither will confidence in the honesty of elections rise, for it would rise only if voters were persuaded that such laws reduce the incidence of such fraud. (9) With the benefit of new survey data, we explore in this Article whether experience with these laws over the past half-decade has changed public beliefs about the incidence of fraud and the tendency to participate in elections. We conclude that there continues to be no empirical evidence that the presence of photo ID laws has a salutary effect on voter confidence. If anything, the evidence we present here suggests that the rise of these laws has coincided with a politicization of opinions about election administration, leading to a slight increase in voter beliefs about the prevalence of fraud.

In revisiting previous empirical research about the relationship between strict voter ID laws, citizen confidence, and voter turnout, we also aim to demonstrate that courts (and many social scientists) have erred by adopting an obsolete theory about how citizens incorporate information about public policy into their attitudes about government. The theory that has underlain analysis of the effects of strict voter identification laws assumes that direct experience with a photo identification requirement will demonstrate to voters that mechanisms are in place to safeguard against voter fraud; seeing or hearing about these laws being passed and implemented will reassure voters about the honesty of the electoral process. Thus, by this argument, even if strict voter ID laws do not actually have a deterrent effect on fraud, and even if impersonation fraud occurs with a frequency that approaches zero, when a voter observes a strict ID law being implemented, she or he will be reassured about the efficacy of elections, and this reassurance redounds to the benefit of democracy.

However, this theory underlying how courts and many scholars have thought about the link in the public mind between voter ID laws and voter reassurance is contrary to the prevailing view within the public opinion literature about how mass publics assess the efficacy of public policy, especially policies that have partisan overtones. The prevailing view is that political partisanship is a deeply held identity among many (and even most) voters, and it is through this identity that they judge government policies. (10) Voters take their cues from party leaders when they judge how well policies are working. (11) If elites associated with the two parties take similar positions on an issue, or if they take no identifiably partisan positions, the mass public will not judge the issue in a partisan manner and is likely to judge it based on factors such as demographics, economic interests, etc. If statements and positions by party elites diverge, opinions by followers will diverge as well, sometimes overshadowing objective interests that the followers themselves might otherwise have. This view has been confirmed across a wide variety of policy domains, both in domestic and foreign policy. (12) There is no reason to believe election policy should be any different. Indeed, given the direct relationship between election regulation and party success at the polls, we should expect partisanship to be an even more powerful predictor of opinion in this domain than in others.

We begin in Part 1 by examining attitudes toward voter identification laws. We find continued support for such laws but a widening gap in support between Democrats and Republicans. The partisan structure of opinion is particularly pronounced among those who follow the news closely and are therefore more likely to pick up signals as to the Democratic and Republican positions on voter identification. Most of the movement in public opinion that we discern from the half-decade of polls has occurred among Democrats, who are now much less likely to support voter identification laws than Republicans. Part II examines attitudes about the prevalence of voter fraud. Here, too, we find increasing divergence between partisans, with Republicans more likely than Democrats to believe that voter impersonation fraud is very common. However, as in our earlier article, we find no relationship between the presence of a voter ID law and respondents' belief in the frequency of fraud. We also do not find any relationship between belief in voter fraud and either reported turnout in a previous election or intention to vote in an upcoming election. Part III briefly discusses survey results concerning why people support voter identification and whether voters know of the existence of restrictive photo identification laws. Although, once again, we find widespread support for such laws, we also find that a substantial share of the population does not know whether or not their state requires photo identification in order to vote. Part IV presents our conclusions, which can be stated succinctly here. Attitudes have changed and become more polarized with respect to voter ID requirements and voter fraud, but we continue to find no relationship between the existence of an ID law and greater voter confidence. This lack of a relationship may be due to the fact that such laws are unevenly enforced or are not salient to a large group of voters. It could also be due, we suspect, to the fact that attitudes about voter fraud have less to do with the precise electoral legal regime in place and its success or failure and more to do with attitudes toward the political system as a whole. In this respect, attitudes about voter fraud are like other measures of political alienation or lack of trust in government: they piggyback onto larger concerns about government competence and specific opinions about the incumbents currently in control.

1. The Changing Partisan Structure of Support for Voter Identification Laws

   We begin our empirical investigation by reviewing support for voter ID laws and the degree to which that support has become associated with partisanship. Recent survey data suggest partisan identity is a powerful variable in predicting attitudes toward voting in general. Democrats are more likely than Republicans, for example, to consider voting to be a right rather than a privilege. Moreover, when forced to choose, Democrats are more likely to say it is more important to make voting easy, rather than more secure. (13) Because most new voting restrictions, such as ID laws, have been passed by Republican-controlled legislatures, (14) we should not be surprised to find that when such policy debates become salient and well known, Republican and Democratic respondents will favor positions consistent with party elites.

   The partisan division on voter identification laws, though it has always been present, has grown in recent years. In a review of support for photo voter ID laws, Gronke et al. examined all publicly available public opinion data about support for strict (i.e., photo) voter ID laws since the mid-2000s. (15) They found that when questions were asked about the matter...