University of Southern California Dissertations and Theses

Can election reforms improve democracy?

(USC Thesis Other)

Can election reforms improve democracy?

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content CAN ELECTION REFORMS IMPROVE DEMOCRACY? by Matthew Nelson A Dissertation Presented to the FACULTY OF THE USC GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (POLITICAL SCIENCE AND INTERNATIONAL RELATIONS) August 2022 Copyright 2022 Matthew Nelson ii Acknowledgments It takes a village to raise a scholar and I wouldn’t have been able to write this dissertation without the extraordinary support and guidance I’ve received. I am extremely grateful for my dissertation committee: Christian Grose, James Lo, and Pamela McCann. Your passion for research and wealth of knowledge have been invaluable. You are not only amazing scholars and mentors, but also amazing people filled with as much kindness as knowledge. Thank you for all your help and advice. I would also like to extend my deepest gratitude to Bob Shrum. You are an amazing mentor, and it was a privilege to be your teaching assistant. Unfortunately, a lot of our time together was spent on Zoom during a pandemic, but it was always a pleasure to talk and learn from you. Thank you for sharing your wisdom and for believing in me. Last but not least, I am nothing without my family and friends. I am so lucky to have y’all. Thank you for giving me the love and laughs that fuel my life. iii Table of Contents Acknowledgments.......................................................................................................................... ii List of Tables................................................................................................................................. iv List of Figures................................................................................................................................. v Abstracts........................................................................................................................................ vi Introduction..................................................................................................................................... 1 Chapter One: Independent Redistricting Commissions are Associated with More Competitive Elections................................................................................................................... 3 Chapter Two: Independent Redistricting Commissions Increase Voter Perceptions of Fairness 12 Chapter Three: Top-Two Primaries Can Moderate Election Outcomes....................................... 25 References..................................................................................................................................... 50 Appendices.................................................................................................................................... 56 iv List of Tables Independent Redistricting Commissions are Associated with More Competitive Elections Table 1: Percent of Independent Commission-Drawn Districts…………………………………. 4 Table 2: Logistic Regression…………………………………………………………………..… 9 Table 3: Odds Ratio Conversion (relative to legislative redistricting)……………………….… 10 Independent Redistricting Commissions Increase Voter Perceptions of Fairness Table 4: Vignettes for Three Experimental Conditions……………………………………...…. 17 Table 5. Independent Commissions Have Greater Perceptions of Fairness……………...…….. 19 Table 6. Independent Commissions Reduce Voter Expectations that Electoral Lines Will be Unfairly Drawn to Favor One Political Party…………………………………...……. 20 Table 7. Independent Commissions Reduce Voter Perceptions that Partisan Gerrymandering is a Problem…………………………………………………………………………..……. 21 Top-Two Primaries Can Moderate Election Outcomes Table 8: Candidate Profiles…………………………………………...………………………… 28 Table 9: Gender Balance………………………………………………………………..………. 30 Table 10: Party Balance………………………………………………...………………………. 30 Table 11: Top-Two Primary: Which candidate did respondents prefer?...................................... 31 Table 12: Republican Closed Primary: Which candidate did respondents prefer?...................... 31 Table 13: Democratic Closed Primary: Which candidate did respondents prefer?...................... 32 Table 14: Do you want your state to use a Top-Two or Closed primary?.................................... 33 Table 15: Candidate A vs Candidate B: Which candidate did respondents prefer?..................... 34 Table 16: Candidate D vs Candidate C: Which candidate did respondents prefer?..................... 34 Table 17: Candidate A vs Candidate C: Which candidate did respondents prefer?..................... 35 Table 18: Candidate A vs Candidate D: Which candidate did respondents prefer?..................... 35 Table 19: Candidate D vs Candidate B: Which candidate did respondents prefer?..................... 35 Table 20: Candidate B vs Candidate C: Which candidate did respondents prefer?..................... 36 Table 21: General Election vote for center-right Republican versus far-right Republican…..… 40 Table 22: General Election vote for center-left Democrat versus far-left Democrat………...… 42 Table 23: General Election vote for center-right Republican versus far-left Democrat……….. 44 Table 24: General Election vote for center-right Republican versus center-left Democrat...….. 45 Table 25: General Election vote for center-left Democrat versus far-right Republican………... 46 Table 26: General Election vote for far-right Republican versus far-left Democrat………...…. 47 v List of Figures Independent Redistricting Commissions are Associated with More Competitive Elections Figure 1: 2018 Redistricting Institutions……………………………………………..………….. 5 Figure 2: 2018 Redistricting Institutions Map…………………………………………..……….. 5 Top-Two Primaries Can Moderate Election Outcomes Figure 3: Election Question, General Election Example- one randomized condition……..…… 29 vi Abstracts Independent Redistricting Commissions are Associated with More Competitive Elections Election competition is crucial for democratic accountability, yet most U.S. House elections are uncompetitive. Using district-level data from 1982 to 2018, I examine the relationship between redistricting institutions and U.S. House election competition. I separate independent and political commissions and find that relative to legislative redistricting, independent commissions are 1.98 times as likely to have competitive elections; and they decrease incumbent party wins by 60%. Independent Redistricting Commissions Increase Voter Perceptions of Fairness Do independent redistricting commissions cause voters to view the redistricting process as fair? Scholars have focused on the role that redistricting plays in political outcomes, but the process built into institutions also signals important values such as fairness. I argue that independent redistricting commissions improve voter attitudes toward fairness in the redistricting process. In two survey experiments, I find that voters are more likely to rate the redistricting process as fair and to believe that partisan gerrymandering is not a problem when they learn about their state’s independent redistricting commission. In contrast, voters in a control group and in a randomized treatment group who learn that state legislators conduct redistricting are less likely to view the redistricting process as fair. Top-Two Primaries Can Moderate Election Outcomes How does the design of primary systems affect voter decision-making and election outcomes? I conduct a survey experiment with participants randomly assigned to vote in either a Top-Two or Closed primary election and in a series of general election scenarios. I find that same-party general elections lead to potential voting coalitions between moderate members of one party and members of the opposite party. In this scenario of same-party general elections, Top-Two primary systems lead voters to choose more moderate candidates. 1 Introduction Can election reforms improve democracy? While activists and the media typically claim that election reforms have huge consequences, scholars tend to emphasize that many reforms may not work as reformers intended, or that the reforms have no impact at all (ACLU 2017; Cantoni and Pons 2019). In contrast to past work in political science, I argue redistricting and primary reforms affect political outcomes. Reform efforts such as independent redistricting institutions and Top-Two primary systems shape the choices available to voters, candidates, and public officials. These institutions also influence people’s attitudes toward government, shaping voters’ views of fairness and accountability. The first essay of my dissertation, Independent Redistricting Commissions are Associated with More Competitive Elections, examines the relationship between redistricting institutions and U.S. House election competition. I argue maps drawn by legislators reduce election competition because electoral motivations and the absence of formal constraints allow the drawing of uncompetitive districts. Instead of legislative redistricting, reformers propose giving redistricting authority to independent commissions composed of citizens. Using district-level election data from 1982 to 2018, I estimate logistic regression models and find that relative to legislative redistricting, independent commissions are 1.98 times as likely to have competitive elections and they decrease incumbent party wins by 60%. The second essay of my dissertation, Independent Redistricting Commissions Increase Voter Perceptions of Fairness, examines how the design of redistricting institutions affects voter attitudes. In two survey experiments, I find that voters are more likely to rate the redistricting process as fair and to believe that partisan gerrymandering is not a problem when they learn about their state’s independent commission. In contrast, voters in a control group and in a 2 randomized treatment group who learn that state legislators conduct redistricting are less likely to view the redistricting process as fair. The third essay of my dissertation, Top-Two Primaries Can Moderate Election Outcomes, examines how the institutional design of primary systems affects voter decision- making and election outcomes. Reformers propose replacing Closed primaries with Top-Two primaries for various reasons. Closed primaries are restricted by party affiliation and guarantee that one Republican and one Democrat proceeds to the general election. On the other hand, Top- Two primaries are open to all voters and can lead to a general election where both candidates are from the same party. I conduct a survey experiment with participants randomly assigned to vote in either a Top-Two or Closed primary election and in a series of general election scenarios. The elections feature hypothetical candidates differentiated by their positions on abortion and immigration. I find that same-party general elections lead to potential voting coalitions between moderate members of one party and members of the opposite party. In this scenario of same- party general elections, Top-Two primary systems lead voters to choose more moderate candidates. In sum, my dissertation shows that election reforms matter. Can election reforms improve democracy? If you think more competitive elections, a fairer redistricting process, and less extreme legislators are good for democracy, then yes. 3 Chapter One: Independent Redistricting Commissions are Associated with More Competitive Elections Partisan motivations for redistricting have existed since the first Congress. George Washington, a Federalist, claimed Virginia districts were “arranged as to place a large proportion of those who are called Antifederalists in that Station” (Hunter 2011). Redistricting now determines the balance of power between Democrats and Republicans, but similar motivations remain. Former Speaker of the House, Newt Gingrich, R-GA., claimed Democrats “get to rip off the public in the states where they control and protect their incumbents, and we get to rip off the public in the states we control and protect our incumbents” (Eilperin 2006). To remedy the conflict of interest apparent in legislative redistricting, reformers have proposed to create independent commissions responsible for redistricting. However, the efficacy of redistricting reform is contested in political science. Some studies indicate that districts drawn by non-legislative institutions are more competitive than districts drawn by state legislatures (Carson 2014; Lindgren and Southwell 2013; McDonald 2006), while other studies find redistricting institutions have little effect on electoral competition (Abramowitz 2006; Forgette 2009; Henderson 2018; Masket 2012). This research extends Carson (2014) by separating independent and political commissions, which were previously grouped together despite different institutional designs. With new data from the 2010 redistricting cycle this analysis is now possible. Existing research on independent commissions may have found null effects because of small sample sizes and a lack of statistical power. Table 1 shows the percentage of districts drawn by independent redistricting commissions in the past four redistricting cycles. In addition to the improvements 4 caused by new data, existing statistical models were improved with state and year fixed effects and clustered standard errors. Table 1: Percent of Independent Commission-Drawn Districts 1982-1990 1992-2000 2002-2010 2012-2018 0.5% 2% 4% 17% I argue maps drawn by politicians reduce election competition because electoral motivations and the absence of formal constraints allow the drawing of uncompetitive districts. Conversely, independent redistricting commissions lead to greater electoral competition because those drawing the lines are motivated by other goals. I analyze U.S. House Elections from 1982- 2018 and find that relative to legislative redistricting, independent commissions are 1.98 times as likely to have competitive elections and decrease incumbent party wins by 60%. Institutional Design and Electoral Competition The United States Constitution mandates each state redraw their congressional districts after the decennial census but does not delegate responsibility to a specific institution. State legislatures draw the majority of congressional districts with district maps created in legislative committees and subject to a vote in both chambers before requiring the governor’s signature. Other states use a political commission to draw their district lines. The design of political commissions varies, with some states allowing the political commission to directly enact district maps and others requiring legislative approval. Independent redistricting commissions separate the legislature from the redistricting process even further by prohibiting legislators from being members. Finally, in some instances, due to legislative inaction or litigation, courts are required 5 to draw district boundaries (Levitt 2019). Figure 1 shows the number of districts different redistricting institutions were responsible for in the 2018 election. In 2018, there were 428 drawn congressional districts and seven states with one at-large congressional district. Figure 2 shows the redistricting institution responsible for 2018 congressional districts. Figure 1: 2018 Congressional Redistricting Institutions Figure 2: 2018 Congressional Redistricting Institutions Map 6 Mapmakers can affect electoral competition by packing minority party voters into safe districts to waste their votes, or by spreading out voters of the minority party to splinter their support (Carson and Crespin 2004). All states are mandated to draw districts as equal in population as possible and prohibited from discriminating on the basis of race, but states are allowed to establish their own additional criteria. Typical redistricting criteria include compactness, contiguity, preservation of communities of interest, and preservation of prior districts. Some states have criteria related to incumbent protection, such as avoiding districts that would create an election between incumbents and prohibiting (or explicitly allowing) mapmakers to favor incumbents. I expect the institutional design of redistricting institutions to affect electoral competition because of the different motivations of members and the constraints placed by rules. Legislators are motivated by their own reelection and the success of their party (Aldrich 1995; Cox and McCubbins 2005; Mayhew 1974). Redistricting can be used to help achieve these goals by creating safe districts where members of the incumbent party are heavily favored to win the election. When redistricting is controlled by the state legislature district maps are treated like normal legislation and biased district maps can be passed by a self-interested partisan majority. I expect state legislative redistricting to be associated with less competitive elections than other types of redistricting institutions. The members of political commissions are either legislators or chosen by legislators, so self-interested motivations are likely. However, unlike legislative redistricting, political commissions are designed to give voice to the minority party. For example, Hawaii’s congressional lines are drawn by a nine-member commission where the majority and minority party leaders in each chamber choose two members and those eight members choose the ninth 7 member (Levitt 2019). In addition to membership rules that countervail partisan motivations, political commissions are constrained by redistricting criteria. In Hawaii, the redistricting commission is not allowed to draw districts that “unduly favor a person or political faction” (Carson and Crespin 2004). I expect political commissions to be associated with more competitive elections than state legislatures. Independent redistricting commissions are designed to function like juries and legislators and party operatives are prohibited from being members. For example, to be eligible to be a commissioner in California, you and your family could not have been a candidate, lobbyist, legislative staff, party employee, or large donor within the last ten years. Furthermore, California’s independent redistricting commission has 5 Democrats, 5 Republicans, and 4 independents to promote impartiality (Levitt 2019). Independent redistricting commissioners may have some partisan motivations, but unlike members of political commissions and state legislatures there are formal rules against running for office in the districts you create. Further, the independent redistricting states all have additional criteria aimed towards creating competitive elections. For example, Arizona’s constitution states “to the extent practicable, competitive districts should be favored where to do so would create no significant detriment to the other goals.” Likewise, commissioners are prohibited from using partisan data when drawing district lines in California (NCSL 2019a; NCSL 2019b). I expect independent redistricting commissions to be associated with more competitive elections than other types of redistricting institutions. Court-administered redistricting gives judges the authority to make new district maps. Judges are expected to apply legal principles neutrally, but existing research shows mapmaking judges have a “mask of neutrality” in which they draw districts that increase competitiveness 8 while advancing their preferred party’s interests (Grose 2011; Peterson 2019). Peterson (2019) finds Democratic judges increase electoral competition by adding Democratic voters to Republican-held districts. I expect court redistricting to be associated with more competitive elections than state legislative redistricting. Empirical Strategy I will analyze redistricting institutions and electoral competition through three district- level variables: incumbent party wins, semi-competitive elections (Democrat’s share of two- party vote between 40% and 60%), and competitive elections (Democrat’s share of two-party vote between 45% and 55%). Election data were retrieved from Gary Jacobson (2020) and include all United States House elections from 1982-2018 for states that conducted redistricting (have multiple districts). Redistricting data were retrieved from Peterson (2019), Levitt (2019), and Ballotpedia (2020). I hypothesize that independent commissions will be associated with less incumbent wins and more semi-competitive and competitive elections than legislative redistricting. Similarly, I hypothesize that political commissions and courts will be associated with less incumbent wins and more semi-competitive and competitive elections than legislative redistricting, but to a lesser degree than independent commissions. To test my hypotheses, I estimate logistic regression models. The independent variables of interest are indicator variables for the different redistricting institutions with legislative-drawn districts as the reference category. Independent Commission is 1 if a district was drawn by an independent redistricting commission and 0 otherwise; Political Commission is 1 if a district was drawn by a political redistricting commission and 0 otherwise; Court is 1 if a district was drawn by a court and 0 otherwise. The dependent variables are different binary measures of electoral competition: Incumbent Party Win is 1 if the incumbent party won the district and 0 otherwise; 9 Semi-Competitive is 1 if a district was semi-competitive and 0 otherwise; Competitive is 1 if a district was competitive and 0 otherwise. The statistical precision of the logistic regression model was improved by using state and year fixed effects and state clustered standard errors (Table 2). State and year fixed effects isolate the relationship between redistricting institutions and electoral competition by controlling for state-to-state and year-to-year differences that are unobserved in the data. Clustered standard errors account for the correlation between observations within each state. The statistical model also includes two control variables that are associated with election competition, but exogenous to the redistricting process. Quality Challenger is 1 if the challenger has held elective office and 0 otherwise. South is 1 if the state was one of the 13 southern states recognized by the Confederacy and 0 otherwise. For the sake of interpretation, the logit coefficients from the model are exponentiated to create odds ratios, which can be interpreted as the odds that an outcome occurs given a particular treatment, compared to the odds of the outcome occurring without that treatment (Table 3). Table 2: Logistic Regression 10 Table 3: Odds Ratio Conversion (relative to legislative redistricting) Results In comparison to states with legislative redistricting, states with independent redistricting commissions have less incumbent party wins, more competitive elections, and more semi-competitive elections (Table 3). The odds ratio indicates independent commissions are 0.40 times less likely to have the incumbent party win, 1.89 times more likely to have semi- competitive races, and 1.98 times more likely to have competitive races (Table 4). These results are statistically significant and robust to various modeling specifications (see Appendices). 1 Political commission-drawn districts have less incumbent party wins and more competitive elections than districts drawn by state legislatures. However, the magnitude of the relationship is less than the one for independent commissions. Relative to legislative-drawn districts, political commission-drawn districts are 0.39 times less likely to have the incumbent party win and 1.93 times more likely to have competitive races (Table 4). Court-drawn districts are 1.43 times more likely to have competitive elections than districts drawn by state legislatures, but there is no statistical difference between court and legislative redistricting in incumbent party success. This result supports existing research that argues courts increase competitiveness while advancing their preferred party’s interests (Peterson 2019). 1 Appendix A includes models: without control variables, fixed effects, and clustered standard errors; with alternate specifications for competitiveness; without uncontested and same-party elections; with data restricted to 2002-2018 11 Conclusion Electoral competition is essential to democratic accountability, yet since the 2010 redistricting cycle the incumbent party has won 94% of congressional elections. I argue that the lack of competitiveness is driven by the design of redistricting institutions. Redistricting conducted by legislators will be influenced by self-interest due to their reelection motivations, while independent commissions remove legislators from the process and place formal constraints on mapmakers. I find a strong, statistically significant relationship between independent redistricting commissions and more competitive elections and less incumbent party wins. California, which completely removes the state legislature from redistricting by using an external body to select commissioners, exemplifies the potential effects of institutional reform. In the decade before independent commissions were implimented only 5.2% of districts in California had a competitive election, since removing redistricting authority from the state legislature the amount of competitive districts has risen to 14.6%. 12 Chapter Two: Independent Redistricting Commissions Increase Voter Perceptions of Fairness Electoral institutions are important not only because they affect the behavior of public officials and representational outcomes (Grose 2011; Riker 1980; Sadhwani and Junn 2018; Thurber and Yoshinaka 2015), but also because they influence voter attitudes (Winburn, Henderson, and Dowling 2017). Voters perceive some institutions are fairer than others, and these perceptions of fairness can strengthen the legitimacy of political institutions (Birch 2008; Collins 2021; Norris 2014; Tolbert, Smith, and Green 2009). In two survey experiments, I find that voters are more likely to rate the redistricting process as fair and to believe that partisan gerrymandering is not a problem when they learn about their state’s independent redistricting commission. In contrast, voters in a control group and in a randomized treatment group who learn that state legislators conduct redistricting are less likely to view the redistricting process as fair. 2 The Redistricting Process Affects Voter Attitudes Toward Fairness Research on the impact of independent commissions on electoral, partisan, and legislator outcomes has been contradictory and mixed. The evaluation of redistricting institutions has led scholars to compare whether independent commissions or legislative-drawn maps: affect the ability for incumbents to win reelection and be responsive to districting changes (Hayes, Hibbing, and Sulkin 2010; Murphy and Yoshinaka 2009); increase electoral competitiveness (Carson, Crespin, and Williamson 2014; Goedert 2016; Masket, Winburn, and Wright 2012; Williamson 2019; Yoshinaka and Murphy 2011); reduce partisan bias and enhance partisan fairness in election outcomes (Best 2021; Cain 2011; McDonald 2004; Stephanopoulos and 2 This essay is part of collaborative work begun in 2020. Additional statistical models and analysis can be found in Grose and Nelson (2021). 13 McGhee 2014); and protect and enhance descriptive representation and fair representational outcomes across racial groups (Casellas 2010; Grose 2006, 2011; Minta and Sinclair-Chapman 2013; Sadhwani and Junn 2018; Tate 2003). In contrast to current emphasis on redistricting’s impact on partisan or electoral outcomes, I argue redistricting institutions also affect how voters perceive the process. Regardless of how the redistricting institution ultimately influences electoral outcomes and legislator behavior, the institution itself can cause voters to feel the process is fair and legitimate. This outcome of perceived fairness among voters is an important theoretical component of democratic institutions (Birch 2008; Rawls 1971; Thompson 2002) and is also one of the key points made by advocates of moving redistricting procedures out of the hands of elected legislators and to independent commissions (Ancheta 2014; Lowenthal 2019). I expect voters who learn the details of an independent commission process to rate that process as fair; and to rate that process as likely to reduce partisan gerrymandering. There are several reasons for this expectation. Voters often value the appearance of neutrality more than processes that appear to favor one political party, such as decisions made by a legislature. While there is some evidence that positive attitudes toward neutral expertise in the United States are weakening (Bayram and Shields 2021), voters have tended to place more trust in institutions that are not legislatures or that are somewhat separated from the political process (Cook and Gronke 2005; Hibbing and Theiss-Morse 1995; McCarthy 2019). For similar reasons, when voters learn that state legislators draw their own district lines, I expect voters to be skeptical about the fairness of the process. Self-interested legislators (Mayhew 1974) draw lines of their electoral boundaries to protect themselves and other incumbents (Cain 1985); and to protect seats of their own political party (McDonald 2004; 14 McKee, Teigen, Turgeon 2006; Yoshinaka and Murphy 2009). A process in which legislators have significant input directly into the process of redrawing their own district lines may raise questions of fairness with voters. In addition, voters already have negative attitudes toward legislatures (Hibbing and Theiss-Morse 1995), and this will be reflected specifically when voters learn about legislative-drawn redistricting procedures. Another reason that voters may perceive independent redistricting commissions as fair is connected to the process by which these commissions came to be. Independent commissions have mostly been passed by voters via ballot propositions. Had they been unpopular, the commissions never would have been passed by voters and subsequently implemented. The popularity of the commission process in states that use them may also lead to more positive attitudes about fairness. Conventional Scholarly Wisdom: Redistricting Institutions Do Not Affect Fairness There is a paucity of scholarship connecting electoral institutions to voter attitudes generally and only one other experimental study examining the impact of redistricting commissions on attitudes toward fairness. In this previous study (Panagopoulos 2013), some respondents were told that district boundaries were going to be drawn by the legislature and some respondents were told they would be drawn by a commission. Then, respondents were asked if they thought the maps would be drawn fair. There was no evidence that independent commissions increased ratings of fairness above legislative-drawn maps. Panagopoulos (2013) is a foundational study and there are several ways I intend to build on its research design. Panagopoulos (2013) was fielded mostly to voters who did not actually live in states with independent commissions. The sample was national, and the idea of an independent commission was merely a hypothetical proposal embedded in a treatment for most 15 respondents in that study. In contrast, I conduct two survey experiments in California, the largest U.S. state that uses an independent redistricting commission. Consequently, respondents in my survey experiments are considering the specific details of their state’s actual redistricting process. California’s independent redistricting commission first drew legislative district lines for the 2012 elections. At the time of the survey fielding, four general elections had been conducted electing U.S. House and state legislative candidates in the first commission-drawn map. A second advantage of my research design is the inclusion of a randomly assigned control group (where no information is provided to the survey respondents) alongside two randomly assigned treatment groups (one describing the California independent redistricting commission process and another suggesting that state legislators redraw the lines). It is important to have a baseline control group that measures generic attitudes toward institutions in general because knowledge of redistricting institutions is low among the populace. Many respondents, in the absence of information for how the institutions or processes work, are likely to have attitudes reflecting low affect and trust in government (Hetherington and Husser 2012; Hibbing and Theiss-Morse 1995). It may be possible that past research finding low respondent ratings of redistricting commissions are simply measuring respondent distaste for American political institutions generally, and not independent redistricting commissions. Third, and perhaps most importantly, very little detail was provided about what redistricting commissions do and what the process is like in Panagopoulos (2013). When voters learn about redistricting commissions in ballot campaigns, the language used to frame the commission versus the legislature is much starker and more extensive. The external validity of my survey experiment is very high because the treatment conditions were written in collaboration with an external group who works in redistricting advocacy. 16 Experiment 1: Registered Voters’ Evaluate Redistricting Commissions Experiment 1 was fielded from February 1 to February 15, 2020 by YouGov as part of the USC Price-USC Schwarzenegger Institute California Issues Poll. YouGov interviewed 1199 respondents who were representative of California’s registered voters. The survey was fielded in both English and Spanish. Respondents were randomly assigned to one of three experimental conditions and then asked questions about their perceptions of the redistricting process. The experimental conditions include an independent commission treatment, a legislative redistricting treatment, and a control condition. The treatments differ in how they describe the redistricting process. The independent commission treatment details how the independent redistricting process works in California, while the legislative treatment describes a redistricting process where legislators draw district lines. The control condition includes no information on how lines are drawn in the state. See Table 4 for the vignettes shown for each experimental condition. 17 Table 4: Vignettes for Three Experimental Conditions Independent Commission Treatment As you may know, congressional and state legislative lines are redrawn every 10 years. In 2021, California is required to redraw the electoral lines for its legislators. In most states, legislators choose their own voters by redrawing their own election boundaries. This means elected legislators pick their voters instead of voters choosing their legislators. In California, we use a different process. A panel of citizens are in charge of redrawing legislators’ lines. This independent panel is called the California Citizens Redistricting Commission, and thousands of California citizens have applied for the Commission this past year. The final commissioners are currently being selected. The commission is chosen by the nonpartisan Bureau of Audits in an attempt to remove politics and partisanship from the process. The citizens commission is composed of equal numbers of Democrats and Republicans, and also has commissioners from other parties or who are registered with no party preference. Once the commission of citizens from different political parties is chosen, they will choose which voters are redrawn into legislators’ election boundaries. This means the independent and nonpartisan citizens commission will redraw the lines that will be used for voters to choose their legislators. How fair is California’s method of redrawing election district lines? State Legislature Treatment As you may know, congressional and state legislative lines are redrawn every 10 years. In 2021, California is required to redraw the electoral lines for its legislators. In most states, legislators choose their own voters by redrawing their own election boundaries. This means elected legislators pick their voters instead of voters choosing their legislators. How fair is California’s method of redrawing election district lines? Control Condition How fair is California’s method of redrawing election district lines? Experiment 1 Results: Registered Voters View Independent Commissions as Fair Following randomization into one of the three experimental conditions, I measured three outcome variables of interest: (1) voters’ attitudes toward the fairness of redistricting in their state; (2) voters’ attitudes towards whether district lines will be unfairly drawn to favor one 18 political party; and (3) voters’ attitudes toward whether partisan gerrymandering was a problem in their state. I present difference of means on these three outcome variables across the two treatment conditions (independent commission treatment and legislature treatment) and the control condition where no additional information was provided. Because respondents are randomly assigned into one of these three conditions, I do not include multivariate regressions with additional right-side variables. All other potential demographic and other characteristics of respondents are “controlled for,” in expectation, due to the randomized experimental design. Later, I consider if the treatment effects are greater conditional on the party identification of the voter. Table 5 displays the results of testing respondent attitudes toward whether the method of redrawing lines is fair. The first dependent variable is measured by asking respondents “How fair is California’s method of redrawing election district lines?” and is coded 1 if the respondent answered fair and 0 if not fair. Voters who learn the details about the process of the independent commission are much more likely to rate the method of redistricting as fair. In this condition, 73.2% of voters rated independent commissions as fair, while only 42.2% said the state’s method of drawing lines was fair in the control group. Learning about how the commission process works in the state to conduct redistricting caused a 31.0 percentage point increase in ratings of fairness among voters (see the difference of means between the independent commission treatment and control conditions in Table 5; p<0.01). Voters perceive a process where the state legislature redraws their own lines as less fair than both the control condition and the independent commission condition. Only 31.6% of voters said state legislative line-drawing was fair (a difference of -10.6 percentage points from the 19 control condition; p<0.01). Taken together, the results in Table 5 reveal that voters are more likely to say that California’s method of redrawing election district lines is fair when given knowledge about California’s independent redistricting commission. Further, voters who were told that legislators control the redistricting process were less likely to say the process is fair than voters who did not receive information about redistricting. Table 5. Independent Commissions Have Greater Perceptions of Fairness Dep. var.: “How fair is California’s method of redrawing election district lines?” 1=Respondent says redistricting method is fair; 0=respondent says not fair Respondents randomly assigned and told… Percentage of respondents saying the way state draws electoral lines is fair Difference of means between treatment and control conditions Independent redistricting commission draws lines 73.2% 31.0** State legislature draws lines 31.6% -10.6** Control group (no information) 42.2% ---------- Means are presented in the 2 nd column, and difference of means are presented in the third column. **p<0.01; *p<0.05. N=1199. Table 6 examines whether voters expect that district lines will be drawn unfairly to favor one political party in their state. The second dependent variable is measured by asking respondents “How likely do you think that California’s election districts will be unfairly drawn to favor one political party?” This outcome variable ranges from 1 to 7 (7=extremely likely; 6=somewhat likely; 5=slightly likely; 4=neither likely nor unlikely; 3=slightly unlikely; 2=somewhat unlikely; 1=extremely unlikely). Table 6 shows that voters in the independent commission treatment are less likely than voters in the legislative treatment and control group to believe one party would unfairly draw the lines. When asked the likelihood of one party drawing unfair lines on a 7-point scale, the independent commission treatment had an average response of 20 4.51, which is between ‘slightly likely’ and ‘neither likely nor unlikely.’ In comparison, the legislative treatment had a mean of 5.22 and the control condition had a mean of 5.12; these values represent the belief that it is ‘slightly likely’ that one political party would unfairly draw lines. The commission treatment is statistically distinct from the control (p<0.01). Table 6. Independent Commissions Reduce Voter Expectations that Electoral Lines Will be Unfairly Drawn to Favor One Political Party Dep. var. “How likely do you think that California’s election districts will be unfairly drawn to favor one political party?” 7= extremely likely; 6=Somewhat likely; 5=Slightly likely; 4=Neither likely nor unlikely; 3=Slightly unlikely; 2=Somewhat unlikely; 1=Extremely unlikely Respondents randomly assigned and told… Mean belief that state will draw electoral lines to favor one party Difference of means between treatment and control conditions Independent redistricting commission draws lines 4.51 -0.61** State legislature draws lines 5.22 0.10 Control group (no information) 5.12 ---------- **p<0.01; *p<0.05. N=1199. Finally, Table 7 provides results for whether voters perceive that partisan gerrymandering is a problem in their state. The third dependent variable is measured with the following question to respondents: “Partisan gerrymandering is when elected legislators draw electoral lines in the legislature that favor their political party. Is partisan gerrymandering a problem in California?” The measure is a 5-point scale with higher values indicating the respondent does not think partisan gerrymandering is a problem (1= Definitely a problem; 2= Somewhat a problem; 3=Neither a problem nor not a problem; 4=Somewhat not a problem; 5=Definitely not a problem). Table 7 shows voters in the independent commission treatment were less likely than voters in the legislative treatment and control group to believe that partisan gerrymandering is a 21 problem in California. Those given the independent commission treatment had a mean response of 2.57, which represents a response in between ‘somewhat a problem’ and ‘neither a problem nor not a problem.’ In comparison, the legislative treatment had a mean of 2.34 and the control condition had a mean of 2.38. These values for the legislative and control conditions also represent a response in between ‘somewhat a problem’ and ‘neither a problem nor not a problem,’ but are closer to ‘somewhat a problem’ than the commission treatment. The independent commission treatment is statistically distinct from the control condition; but the magnitude of the gap is relatively smaller compared to the other two outcome variables. Table 7. Independent Commissions Reduce Voter Perceptions that Partisan Gerrymandering is a Problem Dep. var.: “Partisan gerrymandering is when elected legislators draw electoral lines in the legislature that favor their political party. Is partisan gerrymandering a problem in California?” 1= Definitely a problem; 2= Somewhat a problem; 3=Neither a problem nor not a problem; 4=Somewhat not a problem; 5=Definitely not a problem Respondents randomly assigned and told… Mean response indicating if partisan gerrymandering is problem Difference of means between treatment and control conditions Independent redistricting commission draws lines 2.57 0.19* State legislature draws lines 2.34 -0.04 Control group (no information) 2.38 ---------- **p<0.01; *p<0.05. N=1193. Results Conditional on Political Party of the Respondent I also analyzed the results of Experiment 1 conditional on the partisan identification of the respondent. I did this because we may expect differences by party. Because the Democratic party controlled the state legislature in California when Experiment 1 was fielded, the effect 22 magnitude may be larger for respondents who identify as Republican or independent. Those on the losing end of elections often have attitudes that differ from those that vote more frequently for winning candidates (Blais and Gelineau 2007), so Republican respondents – who often vote for candidates who lose statewide and who lose state legislative elections – may rate a commission more fairly than the legislature if the legislature is associated with the Democratic party. On the other hand, Republicans may be less likely to find the commission fair compared to other voters because the Democratic party has become more identified with institutional reforms such as commissions (Drutman 2020). There are also, though, reasons to think the results will hold for all voters, regardless of party. As can be seen in Appendices B, C, and D, voters in the independent redistricting commission treatment – regardless of political party identification – rate the redistricting process as fairer than those in the control condition. Similarly, Democratic and independent respondents view the commission process as fairer than those in the state legislative treatment condition. The second outcome variable also shows results that are similar among Democratic and Republican voters when examining the commission treatment condition relative to the control condition. There are some mixed and null results on Republican voters on other outcome variables. Experiment 2: Replicating the Results with a Different Respondent Sample I replicated the results of Experiment 1, which was a representative sample of California registered voters in 2020, with a second study. Experiment 2 was fielded via Amazon Mechanical Turk to a convenience sample of 1016 respondents who live in California. Experiment 2 was also fielded in 2020, within a month of Experiment 1. The research design of Experiment 2 is identical to Experiment 1 otherwise. Respondents in Experiment 2 were 23 randomly assigned to the same three conditions with identical language used in Experiment 1, and the same outcome variables were subsequently asked. I present the results from Experiment 2 in Appendix E. The results of Experiment 2 are consistent with and support those found in Experiment 1. Experiment 2 shows that those respondents randomly assigned to learn about the independent commission process of redistricting are more likely to rate it as fair; and this treatment also caused respondents to be less likely to think there is unfairness favoring one political party. The commission treatment also led respondents to be less likely to say that partisan gerrymandering is a problem in California relative to the state legislative treatment, consistent with the results shown above in Experiment 1. I also estimated these three outcomes conditional by party identification of the respondent for Experiment 2, and these results are displayed in Appendices F, G, and H. Across respondents regardless of party in Experiment 2, the independent redistricting commission treatment increases ratings of fairness and decreases perceptions that districts will be unfairly drawn relative to the control group. For the third outcome of whether partisan gerrymandering is a problem, the results are weaker for both Democratic and Republican respondents. These results are generally consistent with the findings from Experiment 1 conditional by respondent party. Conclusion When voters learn about redistricting institutions and processes, it shapes their opinions and attitudes. While political scientists often emphasize the influence (or lack of influence) of electoral institutions on political and policy outcomes, I suggest that institutions themselves are important if they affect the attitudes of voters. In both experiments I conducted, voters are more likely to rate the redistricting process as fair and to say partisan gerrymandering is not a problem 24 when they are randomly assigned to a condition that informs them how the independent commission process works. The electoral reform package before the U.S. House in 2021, H.R.1, included a proposal for all congressional districts to be drawn with independent redistricting commissions. This research shows that voters think these commissions are a fairer process for drawing district lines. 25 Chapter Three: Top-Two Primaries Can Moderate Election Outcomes How does the design of primary systems affect voter decision-making and election outcomes? Closed primaries are restricted by party affiliation and guarantee that one Republican and one Democrat proceeds to the general election. On the other hand, Top-Two primaries are open to all voters and can lead to a general election where both candidates are from the same party. I conduct a survey experiment with participants randomly assigned to vote in either a Top- Two or Closed primary election and in a series of general election scenarios. The elections feature hypothetical candidates differentiated by their positions on abortion and immigration. I find that same-party general elections lead to potential voting coalitions between moderate members of one party and members of the opposite party. In this scenario of same-party general elections, Top-Two primary systems lead voters to choose more moderate candidates. I find that this effect is driven by voters using ideological proximity instead of party cues when the Top- Two presents two general election candidates of the same party. Does the Design of Primaries Change Outcomes? The median voter theorem posits equilibrium occurs when candidates align themselves ideologically with the median voter in a two-candidate election in one dimensional space (Downs 1957). Empirical research has shown this behavior is not demonstrated and that the party primary system may lead candidates to align themselves with the median voter of their party (Green 1994 ; Brady 2007; Nielson and Visalvanich 2017). Some scholars have argued less restrictive primary systems that allow independents to vote have a moderating effect because candidates move to a different, more moderate equilibrium than in a Closed primary (Burden 2004; Hacker and Pierson 2005; Grose 2020; Kaufmann 2003; Gerber and Morton 1998). 26 Meanwhile, other work finds no meaningful differences when you expand the electorate (McGhee and Shor 2017; Ahler 2016; Rogowski and Langella 2015). For example, Ahler (2016) randomly assigned California voters to either a Top-Two or Closed primary election to test whether Top-Two primaries lead to voters choosing the more moderate or ideologically proximate candidate. Voters in the Top-Two primary received the same Top-Two ballot as the actual California election while voters in the Closed primary received a modified ballot that was restricted by partisan affiliation. The authors found no statistical difference between primary systems, but they only conducted an experiment on the primary (first-round) election. I argue that moderating effects are more likely in the general (second-round) election because the Top-Two primary can yield same-party general election contests. Without a party label shortcut, I hypothesize that voters will choose the candidate that is ideologically closest to them. I expect moderate members of the majority party and members of the minority party to be ideologically closer to the candidate that is marginally less extreme. In other words, voters are not necessarily going to vote for the candidate who is ideologically closest to them when partisan affiliation can influence decision-making. Research Design I conduct a survey experiment that isolates the institutional differences between Top-Two and Closed systems. In a Closed primary, voters must be a registered member of the party to vote, independents and members of the opposition party are excluded from participating, and one Republican candidate and one Democratic candidate is guaranteed to proceed to the general election. Alternatively, in a Top-Two primary all voters are allowed to vote, all candidates 27 appear on a common ballot, and the Top-Two candidates regardless of party move on (NCSL 2021). I ask survey respondents questions about their candidate preferences while varying candidate positions on salient issues and randomly varying institutional conditions (Closed system versus Top-Two system). Respondents voted in either a Top-Two or Closed primary (first-round) election and a series of general (second-round) election scenarios. The elections feature different combinations of four candidates. Candidates are differentiated by party affiliation and their views on immigration and abortion (Table 8). Candidate B (far-right) and Candidate C (far-left) are portrayed as ideological extremists and Candidate A (center-right) and Candidate D (center-left) are portrayed as more moderate options. In the Top-Two primary condition, all voters participated and were able to choose any of the four candidates. In the Closed primary condition, the ballot was restricted by party affiliation. Effectively, this design allows for the randomization of the system type; and the randomization of the candidate scenarios (by ideology and party). 28 Table 8: Candidate Profiles Statement on Immigration Statement on Abortion Ideology Candidate A (Republican) American workers are hurt by immigration. We should limit foreign visas and increase security at the border. Life begins at conception. Roe v. Wade should be overturned, and abortion laws should be decided by individual states. Center-Right Candidate B (Republican) Immigrants are dangerous and are stealing American jobs. We should build a wall and increase deportations of illegal immigrants. Abortion is murder. Roe v. Wade should be overturned, and abortion should be banned in the Constitution. Far-Right Candidate C (Democrat) No human being is illegal. We should give citizenship to immigrants living here illegally and expand refugee programs. Abortion is a women's health decision. Roe v. Wade should be protected, and government healthcare should cover abortions. Far-Left Candidate D (Democrat) America is a country of immigrants. We should establish a path to citizenship for immigrants living here illegally. Abortions should be rare, safe, and legal. Roe v. Wade should be protected. Center-Left After the primary elections, participants were randomly assigned to vote in the following general elections: Democrat versus Democrat (far-left vs center-left), Republican vs Republican (far-right vs center-right), Republican versus Democrat (far-right vs far-left), Republican versus Democrat (center-right vs far-left), Republican vs Democrat (left vs far-right), and Republican vs Democrat (center-left vs right). The sequence of general elections and the order of choices were both randomized. Figure 3 shows how a general election was presented to the survey respondents using the online platform Qualtrics. 29 Figure 3: Election Question, General Election example- one randomized condition Voters were asked to place themselves, Donald Trump, Joe Biden, and the candidates from the experiment on a liberal-conservative ideological scale. I expect analysis of these data to show that the design of primary institutions affects voter decision making and electoral outcomes. In both Closed and Top-Two primary elections, I expect voters to choose the candidate from their party that is perceived to be ideologically closest to them. In same-party general elections, where partisan label cannot determine decision-making, I expect voters use ideological proximity to make decisions. On the other hand, in Democrat-versus-Republican general elections, I expect voters to choose the candidate from their party, even if they are 30 perceived to be ideologically farther from them. These hypotheses are one-directional and will be statistically tested with p-values less than 0.05 the threshold for statistical significance. Representativeness and Balance Across Co-variates in the Experiment The survey experiment was fielded to 1598 respondents on Lucid Theorem. This study was formally exempted by USC IRB (study ID UP 21 00682) and preregistered on EGAP prior to fielding the survey. Lucid Theorem collects a nationally representative sample balanced on age 18+, gender, ethnicity, and region. Respondents in the Top-Two and Closed conditions were balanced on age, gender, ethnicity, education, and political party. Table 9 and Table 10 highlight the balance between treatment conditions. The balance shown on these co-variates across conditions is expected, and indicates the randomization was done correctly. Additional demographic information can be found in Appendix I. Table 9: Gender Balance Female Male Non-binary/ third gender Top-Two 55% 44% 1% Closed 52% 47% 1% Table 10: Party Balance Democrat Independent Republican Third Party Top-Two 39% 31% 28% 2% Closed 39% 29% 29% 3% 31 Impact of Randomized Primary Type on Respondent Vote Choice in Primary Elections Voters in the Top-Two primary condition were assigned to a Top-Two primary election with all four candidates on the ballot. Table 11 shows the results of the Top-Two primary election. Out of respondents randomly assigned to the Top-Two condition: 25% chose the center- right Republican (Candidate A), 17% chose the far-right Republican (Candidate B), 26% chose the far-left Democrat (Candidate C), and 32% chose the center-left Democrat (Candidate D). Candidate C and Candidate D were the top two vote-getters in this primary. Both candidates are Democrats, meaning this primary election would lead to a same party general election. Table 11: Top-Two Primary: Which candidate did respondents prefer? Candidate A (Republican) Candidate B (Republican) Candidate C (Democrat) Candidate D (Democrat) Center-Right Far-Right Far-Left Center-Left 25% 17% 26% 32% Voters in the Closed primary condition were assigned to their party’s primary election. Table 12 shows how Republicans randomly assigned to the Closed primary condition voted: 63% chose the center-right Republican (Candidate A) and 37% chose the far-right Republican (Candidate B). The top vote-getter in the Republican primary election, the center-right Republican (Candidate A), would progress to the general election to face the winner of the Democratic primary. Table 12: Republican Closed Primary: Which candidate did respondents prefer? Candidate A (Republican) Candidate B (Republican) Center-Right Far-Right 63% 37% 32 Table 13 shows how Democrats randomly assigned to the Closed primary voted: 55% chose the far-left Democrat (Candidate C) and 45% chose the center-left Democrat (Candidate D). The top vote-getter in the Democratic primary election, the far-left Democrat (Candidate C), would progress to the general election. Independents randomly assigned to the Closed primary condition were told they are unable to vote in Closed primary elections. Table 13: Democratic Closed Primary: Which candidate did respondents prefer? Candidate C (Democrat) Candidate D (Democrat) Far-Left Center-Left 55% 45% Do Voters Treated with a Top-Two or Closed Primary Prefer Top-Two or Closed? While my primary research question seeks to assess how primary systems constrain and shape vote choices, I also wanted to measure attitudes toward the primary system itself. There has been limited research assessing attitudes toward primary systems (except Alvarez and Sinclair 2015). Given this, in addition to the above outcome variables on vote choice, I also asked the following to respondents after they were randomly assigned to vote in the Closed or Top-Two primary election: In a Top-Two primary, Democrats and Republicans run against each other, and the two top candidates advance to the general election. This means that two candidates from the same party can advance to the general election. In a Closed primary, Democrat and Republican voters select their party’s candidate for the general election, and the top candidate from each party advances to the general election. Independent and third party voters are not allowed to vote in Closed primary elections. Do you want your state to use a Top-Two or Closed primary? Table 14 shows how responses varied based on their experience voting in a Top-Two or Closed primary. Out of respondents who voted in the Top-Two primary: 62% preferred the Top- Two system and only 38% preferred the Closed system. Out of respondents who voted in the Closed primary: 56% preferred the Top-Two system and 44% preferred the Closed system. 33 While respondents in both conditions preferred Top-Two, respondents randomly assigned to vote in the Top-Two primary were more likely to want a Top-Two primary than respondents randomly assigned to vote in the Closed primary. The difference of means was statistically significant with a p-value of 0.015. This suggests the experience of voting in a Top-Two primary increases voter’s desire for the system. Table 14: Do you want your state to use a Top-Two or Closed primary? Top-Two Closed Voted in Top-Two 62% 38% Voted in Closed 56% 44% Impact of Randomized Primary Type on Respondent Vote Choice in General Elections In this section, I present the results from the different general election scenarios. As discussed before, my study differs from past work as I randomly assign the type of primary system to respondents; and then ask about vote preferences in both a primary election and then in a series of general elections. Past work has only examined voter choices in survey experiments with primary elections. While same-party general elections are only possible in Top-Two primary systems, respondents in both experimental conditions were randomly assigned to vote in all six general election scenarios. Respondent preferences were consistent in the Top-Two and Closed conditions (Appendix J), which means the experience of voting in a Top-Two primary (first- round) does not change voter attitudes towards candidates. However, Top-Two still affects voter decision making when it constrains general election choices to only one party. Then, differences in candidate preferences by party affiliation build coalitions of voters across parties. 34 Table 15 shows the results of the same-party general election between a center-right Republican (Candidate A) and a far-right Republican (Candidate B). The center-right Republican won receiving 70% of all votes. Both Democrats and Republicans preferred the center-right candidate, but support for the far-right candidate was 9% higher among Republicans. Table 15: Candidate A vs Candidate B: Which candidate did respondents prefer? Candidate A (Republican) Candidate B (Republican) Center-Right Far-Right All Voters 70% 30% Democratic Voters 75% 25% Republican Voters 66% 34% Table 16 shows the results of the same-party general election between a center-left Democrat (Candidate D) and a far-left Democrat (Candidate C). The center-left Democrat won receiving 62% of all votes. Both Democrats and Republicans preferred the center-left candidate, but support for the far-left candidate was 19% higher among Democrats. Table 16: Candidate D vs Candidate C: Which candidate did respondents prefer? Candidate D (Democrat) Candidate C (Democrat) Center-Left Far-Left All Voters 62% 38% Democratic Voters 55% 45% Republican Voters 74% 26% Voters overwhelmingly chose the candidate that shared their party identification in all Democrat-Republican general elections. Table 17 shows the results of a Democrat-Republican general election between a center-right Republican (Candidate A) and a far-left Democrat (Candidate C). The far-left Democrat won receiving 52% of all votes. 75% of Democrats selected the far-left Democrat and 77% of Republicans selected the center-right Republican. 35 Table 17: Candidate A vs Candidate C: Which candidate did respondents prefer? Candidate A (Republican) Candidate C (Democrat) Center-Right Far-Left All Voters 48% 52% Democratic Voters 25% 75% Republican Voters 77% 23% Table 18 shows the results of a Democrat-Republican general election between a center- right Republican (Candidate A) and a center-left Democrat (Candidate D). The center-left Democrat won receiving 57% of all votes. 77% of Democrats selected the center-left Democrat and 70% of Republicans selected the center-right Republican. Table 18: Candidate A vs Candidate D: Which candidate did respondents prefer? Candidate A (Republican) Candidate D (Democrat) Center-Right Center-Left All Voters 43% 57% Democratic Voters 23% 77% Republican Voters 70% 30% Table 19 shows the results of a Democrat-Republican general election between a center- left Democrat (Candidate D) and a far-right Republican (Candidate B). The center-left Democrat won receiving 62% of all votes. 77% of Democrats selected the center-left Democrat and 62% of Republicans selected the far-right Republican. Table 19: Candidate D vs Candidate B: Which candidate did respondents prefer? Candidate D (Democrat) Candidate B (Republican) Center-Left Far-Right All Voters 62% 38% Democratic Voters 79% 21% Republican Voters 38% 62% 36 Table 20 shows the results of a Democrat-Republican general election between a far-right Republican (Candidate B) and a far-left Democrat (Candidate C). The far-left Democrat won receiving 58% of all votes. 79% of Democrats selected the far-left Democrat and 70% of Republicans selected the far-right Republican. Table 20: Candidate B vs Candidate C: Which candidate did respondents prefer? Candidate B (Republican) Candidate C (Democrat) Far-Right Far-Left All Voters 42% 58% Democratic Voters 21% 79% Republican Voters 70% 30% The general election results demonstrate how same-party general elections in Top-Two primary systems can lead to more moderate candidates, but this effect is only observed in same- party general elections. The general (second-round) election that would have followed the Closed primary (first-round) election was between a center-right Republican and a far-left Democrat (Table 17). Since voters mostly selected the candidate that shared their party in a Democrat-versus-Republican general election, the far-left Democrat won. The general (second- round) election that would have followed the Top-Two primary (first-round) election was between a center-left Democrat and a far-left Democrat (Table 16). A coalition of Democratic and Republican voters selected the more moderate candidate and the center-left Democrat won when the respondents faced a same-party general election. These results suggest that same-party general elections in the Top-Two system produce different effects than different-party general elections (which always happen in a Closed system, but sometimes also happen in a Top-Two system). 37 Vote Choice by Ideology or by Party in General Elections The above analyses rely simply upon the aggregated general election results within each randomized condition. While this provides important evidence suggesting moderate candidates do well in same-party general elections and party cues are most important in different-party general elections, I still want to probe more regarding the relationship of individual voter choices. Individual voters will be located in different parts of the ideological space. Thus, I want to see if voters choose the spatially closer candidate in all of the randomized conditions, or if voters choose the spatially closer candidate only in the randomized condition with Top-Two and same-party general election candidates. Theoretically, I would expect that voters faced with a Closed system that features a Democrat versus a Republican in the general election; or a Top-Two system that resulted in a Democrat versus a Republican in the general election would rely on party cues in making decisions. In these different-party general election contexts, the ideological position of the candidate vis-à-vis the respondent’s ideology will not matter as much and many voters will simply vote based on shared partisanship with the candidate. In the same-party general election condition, though, I would anticipate that ideology of the candidate vis-à-vis the respondent will matter more. In same-party contexts, voters will be more likely to choose the ideologically proximate candidate. To test this, I need to create statistical models that incorporate information about the candidate party, ideology (far-left, center-left, center-right, or far-right) and the voter’s ideology. Empirically, this is also important as the Lucid Theorem sample had a somewhat uneven distribution of respondents by ideology (more far-left respondents, for instance). I created linear probability models to determine whether the effect of the randomized Top-Two and Closed systems are conditional on respondent ideology or partisan affiliation. 38 Vote for Candidate = β0 + β1Top-Two + β2Closer Ideology + β3Same Party + β4Ideology x Party + ε The dependent variable is a binary indicator of vote choice for a particular candidate. Vote for Candidate is coded one if the respondent voted for the candidate, and zero otherwise. For sake of interpretation, I focus on vote choice for the moderate candidates. There are four binary independent variables. Top- Two is one if the respondent was in the Top-Two primary condition, and zero otherwise (for the Closed condition). Closer Ideology is one if the respondent was ideologically closer to the candidate, and zero otherwise. The Closer Ideology measure was constructed from survey respondents’ answers to questions about their ideological position and the ideological position of candidates on a 7-point liberal-conservative scale. Same Party is one if the respondent had the same party affiliation as the candidate, and zero otherwise. Ideology x Party is the interaction between Closer Ideology and Same Party. Coefficients can be interpreted as the change in probability of voting for a candidate when the value of the variable goes from zero to one. It is important to remember that the Top-Two variable was randomized to the voter at the primary stage, so this variable may not have a statistical effect on vote choices in the general election. However, the impact of the Top-Two system may be observed in the randomized candidate conditions (where candidates of different ideological backgrounds are randomized to be in the same party; or to be in a different party). Thus, in the analyses below, a null effect on the Top-Two variable does not imply the Top-Two system has no impact on general election vote choices. It simply means randomization into that condition in the primary election stage of the survey has no effect. The effects found while varying the parties and ideological positions of 39 the candidates also measure the impact of a feature of the Top-Two system (general elections with candidates from the same party). Table 21 shows the result of the ordinary least squares regression for the center-right Republican versus far-right Republican election. The coefficient for Same Party was negative and statistically significant. The coefficient for Ideology x Party was positive and statistically significant. The predicted value of voting for the center-right Republican for respondents who were in the Top-Two condition, were ideologically closer, and had the same party affiliation was 0.76. The predicted value for respondents who were in the Top-Two condition, were ideologically further, and had the same party affiliation was 0.61. The predicted value for respondents who were in the Top-Two condition, were ideologically closer, and had a different party affiliation was 0.74. The predicted value for respondents who were in the Top-Two condition, were ideologically further, and had a different party affiliation was 0.76. This suggests that Democrats – who did not share either candidates’ party affiliations as both candidates were Republicans – were effectively not motivated by being closer ideologically. However, among Republicans, when faced with a same-party Republican-versus-Republican choice, there was a large magnitude of the effect of ideology (GOP voters were 15 percentage points more likely, all else equal, to vote for the ideologically proximate same-party candidate). These predicted values were calculated by varying the underlying values of each independent variable and multiplying them by the coefficients. They show that one of the features of the Top-Two system leads voters to make ideological choices. The variable for Ideology x Party is a significant predictor of vote choice. The impact of party will be substantially muted, presumably, when the Top-Two randomized condition produced a same- party general election. In this instance, being ideologically closer to the candidate will influence 40 voter choices, but it is conditional on whether the voter shares the party of the candidate. Republican voters consider ideological proximity when both general-election candidates are Republicans, but the impact of ideological proximity on Democrats in the same scenario is substantially muted. Table 21: General Election vote for center-right Republican versus far-right Republican Table 22 shows the result of the regression for the center-left Democrat versus far-left Democrat election. Here again, I am examining the role of party and ideology on vote choices within a scenario that is only possible in the Top-two primary condition – two Democrats running against one another. The coefficient for Same Party was negative and statistically significant. The predicted value of voting for the center-left Democrat for respondents who were in the Top-Two condition, were ideologically closer, and had the same party affiliation was 0.65. The predicted value for respondents who were in the Top-Two condition, were ideologically further, and had the same party affiliation was 0.49. This means that for Democratic respondents, 41 when faced with the choice of two Democratic candidates, ideological proximity between voter and candidate has a substantively meaningful effect. Democratic voters are 16 percentage points more likely to vote for the ideologically closer candidate when the general election choices are two Democrats. The results in Table 22 also show that ideological proximity is meaningful for Republican voters when faced with two Democratic candidates in the general election, but the substantive effect size is much smaller than found with Democratic respondents facing two Democratic candidates. The predicted value for respondents who were in the Top-Two condition, were ideologically closer, and had a different party affiliation was 0.79. The predicted value for respondents who were in the Top-Two condition, were ideologically further, and had a different party affiliation was 0.72. Thus, Republican voters are more 7 percentage points more likely to choose the ideologically closer Democratic candidates when choosing between two Democratic candidates. However, the effect size of ideological closeness is much smaller than when a Democratic voter is choosing between two Democratic general election candidates. 42 Table 22: General Election vote for center-left Democrat versus far-left Democrat Next, I turn to estimating the impact of ideology and shared party on respondents’ vote choices under conditions in which the two general-election candidates have different parties. Table 23 shows the result of the regression for the center-right Republican versus far-left Democrat election. The coefficient for Closer Ideology was negative and statistically significant. The coefficient for Same Party was positive and statistically significant. The coefficient for Ideology x Party was positive and statistically significant. The predicted value of voting for the center-right Republican for respondents who were in the Top-Two condition, were ideologically closer, and had the same party affiliation was 0.80. The predicted value for respondents who were in the Top-Two condition, were ideologically further, and had the same party affiliation was 0.73. The predicted value for respondents who were in the Top-Two condition, were ideologically closer, and had a different party affiliation was 0.20. The predicted value for 43 respondents who were in the Top-Two condition, were ideologically further, and had a different party affiliation was 0.28. These results from Table 23 show that shared party affiliation between voter and candidate is much more influential than being closer ideologically. For Republicans, 73% voted for the Republican candidate even though the Democratic candidate was ideologically closer; while 80% of Republicans who were ideologically closer to the Republican candidate and not to the Democratic candidate voted for the Republican candidate. Among Democratic respondents, only 20% voted for the Republican candidate when the Democratic candidate was ideologically closer; and only 28% of Democratic respondents voted for the Republican candidate even if the Republican candidate was ideologically closer to the respondent than was the Democratic candidate. Ideology matters around the margins in different-party general elections, but respondents rely much more on the party cue when faced with candidates of different political parties. Such a scenario can occur in a Closed system (in all general elections) and may also occur sometimes (but not always) in the Top-Two system’s general elections. 44 Table 23: General Election vote for center-right Republican versus far-left Democrat Similar results are found under different-party-candidate general elections with different ideological profiles of candidates. Table 24 shows the result of the regression for the center-right Republican versus center-left Democrat election. The coefficient for Same Party was positive and statistically significant. The coefficient for Ideology x Party was positive and statistically significant. The predicted value of voting for the center-right Republican for respondents who were in the Top-Two condition, were ideologically closer, and had the same party affiliation was 0.77. The predicted value for respondents who were in the Top-Two condition, were ideologically further, and had the same party affiliation was 0.65. The predicted value for respondents who were in the Top-Two condition, were ideologically closer, and had a different party affiliation was 0.24. The predicted value for respondents who were in the Top-Two condition, were ideologically further, and had a different party affiliation was 0.23. Again, in this general election with a center-right Republican and a center-left Democrat, voters are much more 45 likely to use shared partisanship with a candidate as a cue in casting a ballot than ideological proximity to the candidate. Table 24: General Election vote for center-right Republican versus center-left Democrat Table 25 shows the result of the regression for the center-left Democrat versus far-right Republican election. The coefficient for Closer Ideology was positive and statistically significant. The coefficient for Same Party was positive and statistically significant. The predicted value of voting for the center-left Democrat for respondents who were in the Top-Two condition, were ideologically closer, and had the same party affiliation was 0.86. The predicted value for respondents who were in the Top-Two condition, were ideologically further, and had the same party affiliation was 0.76. The predicted value for respondents who were in the Top- Two condition, were ideologically closer, and had a different party affiliation was 0.49. The predicted value for respondents who were in the Top-Two condition, were ideologically further, 46 and had a different party affiliation was 0.34. In this scenario, both shared party between respondent and candidate; and ideological proximity are important. Yet again, though, the magnitude of shared party affiliation is larger than ideological closeness for respondents. Table 25: General Election vote for center-left Democrat versus far-right Republican Table 26 shows the result of the regression for the far-right Republican versus far-left Democrat election. The coefficient for Closer Ideology was negative and statistically significant. The coefficient for Same Party was positive and statistically significant. The coefficient for Ideology x Party was positive and statistically significant. The predicted value of voting for the far-right Republican for respondents who were in the Top-Two condition, were ideologically closer, and had the same party affiliation was 0.79. The predicted value for respondents who were in the Top-Two condition, were ideologically further, and had the same party affiliation was 0.57. The predicted value for respondents who were in the Top-Two condition, were ideologically closer, and had a different party affiliation was 0.13. The predicted value for 47 respondents who were in the Top-Two condition, were ideologically further, and had a different party affiliation was 0.24. Again, in this scenario with two candidates from different parties on the general election ballot, voters used the party affiliation of the candidate much more in making their decisions than ideological proximity – even though both factors seem to be associated with vote choice. Table 26: General Election vote for far-right Republican versus far-left Democrat The regression models reveal several patterns for voter decision making. In all the Democrat-Republican elections the coefficient for Same Party was positive, which means sharing the same party affiliation as the candidate increased the probability of voting for a candidate. On the other hand, in both same-party general elections the coefficient for Same Party was negative. To really understand the substantive implications of the results, it is important to 48 consider whether the election featured two candidates of the same party, the party of the respondent and whether it was shared with the candidate, the ideological closeness of the voter, and the interaction of the two. In same-party general elections contests featuring two Democrats, Democratic voters relied upon ideological closeness and voted for the Democrat more proximate. In same-party general election contests featuring two Republicans, Republican voters also relied upon ideological proximity and chose the Republican who was ideologically closer. Such general election outcomes can only occur in a Top-Two system, and thus the removal of a party cue caused voters who shared the party of both candidates to make vote choices based on ideological proximity. This is a goal of reformers who prefer Top-Two primary systems– to move vote choices toward issue or ideological agreement and not simply picking based on shared partisanship between voters and candidates. Interestingly, the magnitude of the effect of ideology was smaller for Democrats who were forced to choose between two Republican candidates; and for Republicans who were forced to choose between two Democratic candidates. Thus, the impact of a Top-Two system on forcing ideological proximity as a voters’ decision rule is greater among those who share the party of the candidates. Democratic voters choosing between a center-right Republican and far-right Republican, for instance, were not particularly moved by being closer to one candidate relative to the other. Since vote choice in these elections was measured as the preference for the moderate candidate, this indicates voters who shared party affiliation with candidates were less likely to select the moderate candidate and voters who did not share party affiliation with either candidate were more likely to select the moderate. This result appears to be due to the ideological location of the voters vis-à-vis the candidates in same-party general election scenarios. 49 In contrast, different-party general elections led voters to make choices that relied heavily on the party affiliation of a candidate, and ideological proximity had a much smaller magnitude. These types of general elections always occur in a Closed system, but also sometimes occur in a Top-Two system as well. Further, in elections between the far-left Democrat and the center-right and far-right Republicans, the coefficient on Closer Ideology was negative. This suggests voters were selecting based on party affiliation in these elections, regardless of their perceived ideological distance from the candidates. Conclusion In comparison to restrictive, Closed primary systems, Top-Two primary systems allow more voters to participate in the primary election and theoretically should lead to more moderate candidates. Instead of conducting an observational study, I conduct an experiment that directly tests whether Top-Two systems lead to more moderate candidates. I find that same-party general elections create coalitions between moderate members of one party and members of the opposite party. This effect seems driven by voters who must use ideological proximity as a decision rule when the Top-Two presents two general election candidates of the same party. In this scenario, Top-Two primary systems lead to more moderate electoral outcomes. Reformers interested in promoting bipartisanship should advocate for primary system reform. Importantly, I find that a majority of voters prefer Top-Two primaries over Closed primaries, and the act of participating in a Top-Two system improves approval for the system relative to Closed systems. 50 References Abramowitz, Alan, Brad Alexander, and Matthew Gunning. 2006. “Incumbency, Redistricting,and the Decline of Competition in US House Elections.” Journal of Politics 68 (1): 75–88. Ahler, Douglas J., Citrin, Jack, and Lenz, Gabriel S. “Do Open Primaries Improve Representation? An Experimental Test of California’s 2012 Top-Two Primary.” Legislative Studies Quarterly 41, no. 2 (May2016): 237–268. Aldrich, John H. 1995. Why Parties?: The Origin and Transformation of Political Parties in America. Chicago: The University of Chicago Press, 1995. Alvarez, R. Michael, and J. Andrew Sinclair. Nonpartisan Primary Election Reform : Mitigating Mischief. Cambridge: Cambridge University Press, 2015. Ancheta, Angelo N. 2014. “Redistricting Reform and the California Citizens Redistricting Commission.” Harvard Law & Policy Review 8:109-140. Bayram, A. Burcu and Todd Shields. 2021. “Who Trusts the WHO? Heuristics and Americans’ Trust in the World Health Organization During the COVID-19 Pandemic.” Social Science Quarterly. https://doi.org/10.1111/ssqu.12977. Best, Robin E, Steve B Lem, Daniel B Magleby, and Michael D McDonald.2021. “Do Redistricting Commissions Avoid Partisan Gerrymanders?” American Politics Research. Birch, Sarah. 2008. “Electoral Institutions and Popular Confidence in Electoral Processes: A Cross-national Analysis.” Electoral Studies 27:305-20. Blais, Andre and Francois Gelineau. 2007. “Winning, Losing, and Satisfaction with Democracy.” Political Studies 55:425-41. Brady, David W., Han, Hahrie, and Pope, Jeremy C. 2007. “Primary Elections and Candidate Ideology: Out of Step with the Primary Electorate?” Legislative Studies Quarterly. Burden, Barry C. 2004. “Candidate Positioning in U.S. Congressional Elections.” British Journal of Political Science 34: 211–27. Cain Bruce E. 1985. “Assessing the Partisan Effects of Redistricting.” American Political Science Review 79:320-33. Cain, Bruce E. 2011. “Redistricting Commissions: A Better Political Buffer?” Yale Law Journal 1808-1845. Carson, Jamie L., and Michael H. Crespin. 2004. “The Effect of State Redistricting Methods on Electoral Competition in United States House of Representatives Races.” State Politics & Policy Quarterly 4:455–69. 51 Carson, Jamie, Michael Crespin, and Ryan Williamson. 2014. “Reevaluating the Effects of Redistricting on Electoral Competition, 1972–2012.” State Politics and Policy Quarterly 14 (2): 165–77. Cantoni, Enrico, and Vincent Pons. “Strict Id Laws Don’t Stop Voters: Evidence from a U.S. Nationwide Panel, 2008-2018.” The Quarterly journal of economics 136, no. 4 (2021): 2615–2660. Casellas, Jason. 2010. Latino Representation in State Houses and Congress. New York: Cambridge University Press. Collins, Jonathan. 2021. “Does the Meeting Style Matter? The Effects of Exposure to Participatory and Deliberative School Board Meetings.” American Political Science Review. DOI: doi.org/10.1017/S0003055421000320. Cook, Timothy E. and Paul Gronke. 2005. “The Skeptical American: Revisiting the Meanings of Trust in Government and Confidence in Institutions.” Journal of Politics 67:784-803. Cox, Gary W, and Mathew D. McCubbins. 2005. Setting the Agenda: Responsible Party Government in the U.S. House of Representatives. Cambridge: Cambridge University Press, 2005. Downs, A. (1957). An economic theory of political action in a democracy. The journal of political economy, pages 135–150. Drutman, Lee. 2020. Breaking the Two-party Doom Loop. Oxford: Oxford University Press. Edwards, Barry, Crespin, Michael, Williamson, Ryan D., and Palmer, Maxwell. 2017. “Institutional Control of Redistricting and the Geography of Representation.” Journal of Politics 79 (2): 722–726. Eilperin, Juliet. 2006. “The Gerrymander That Ate America.” April 17. Retrieved from https://slate.com/news-and-politics/2006/04/this-is-the-only-way-to-fix-congress.html Forgette, Richard, Andrew Garner, and John Winkle. 2009. “Do Re-districting Principles and Practices Affect US State Legislative Electoral Competition?” State Politics and Policy Quarterly 9 (2):151–75. Gerber, Elisabeth R., and Rebecca B. Morton. 1998. “Primary Election Systems and Representation.”Journal of Law, Economics, and Organization 14(1): 304–24. Goedert, Nicholas. 2016. “The Pseudoparadox of Partisan Mapmaking and Congressional Competition.” State Politics and Policy Quarterly 17:47-75. Green, D. P., Shapiro, I., and Shapiro, I. (1994). Pathologies of rational choice theory: A critique of applications in political science. Cambridge Univ Press. 52 Grose, Christian R. 2006. “Bridging the Divide: Interethnic Cooperation; Minority Media Outlets; and the Coverage of Latino, African-American, and Asian-American Members of Congress.” Harvard International Journal of Press/Politics 11:115-30. Grose, Christian R. 2011. Congress in Black and White: Race and Representation in Washington and at Home. New York: Cambridge University Press. Grose, Christian R. and Matthew Nelson. 2021. “Independent Redistricting Commissions Increase Voter Perceptions of Fairness”. Working paper. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3865702 Grose, Christian. (2020). “Reducing Legislative Polarization:Top-Two and Open Primaries Are Associated with More Moderate Legislators”. Journal of Political Institutions and Political Economy. Hacker, Jacob S., and Paul Pierson. 2005. “Abandoning the Middle: The Bush Tax Cuts and the Limits of Democratic Control.” Perspectives on Politics 3: 33–53. Hayes, Matthew, Matthew V. Hibbing, and Tracy Sulkin. 2010. “Redistricting, Responsiveness, and Issue Attention.” Legislative Studies Quarterly 35:91-115. Henderson, John A., Brian T. Hamel, and Aaron M. Goldzimer. 2018. “Gerrymandering Incumbency: Does Nonpartisan Redistricting Increase Electoral Competition?” Journal of Politics 80:1011-16. Hetherington, Marc J. and Jason A. Husser. 2012. “How Trust Matters.” American Journal of Political Science 56:312-25. Hibbing, John R. and Elizabeth Thiess-Morse. 1995. Congress as Public Enemy: Public Attitudes Toward American Political Institutions. New York: Cambridge University Press. Hunter, Thomas Rogers. “The First Gerrymander? Patrick Henry, James Madison, James Monroe, and Virginia’s 1788 Congressional Districting.” Early American studies 9, no. 3 (2011): 781–820. Jacobson, Gary. 2015. “It’s Nothing Personal: The Decline of the Incumbency Advantage in US House Elections.” Journal of Politics 77 (3): 861–873. Jacobson, Gary. 2020. “U.S. House 1946-2018 Election Data”. (May 20, 2020). Distributed by Gary Jacobson. Kaufmann, Karen M., James G. Gimpel, and Adam H. Hoffman. “A Promise Fulfilled? Open Primaries and Representation.” Journal of Politics 65, no. 2 (2003): 457-76. Lemongello, Steven. (September 10, 2020). “Amendment 3: Would a ‘top 2′ primary open up elections or dilute the minority vote?” Orlando Sentinel. 53 Levitt, Justin. 2019. All About Redistricting. Retrieved from http://redistricting.lls.edu/ Lindgren, Eric, and Priscilla Southwell. 2013. “The Effect of Redistricting Commissions on Electoral Competitiveness in US House Elections, 2002–2010.” Journal of Politics and Law 6 (2): 13–18. Lo, James. 2013. “Legislative Responsiveness to Gerrymandering: Evidence from the 2003 Texas Redistricting.” Quarterly Journal of Political Science 8:75-92. Lowenthal, Alan S. 2019. “The Ills of Gerrymandering and Independent Redistricting Commissions as the Solution.” Harvard Journal on Legislation 56:1-22. Masket, Seth E., Jonathan Winburn, and Gerald C. Wright. 2012. “The Gerrymanders are Coming!” PS: Political Science and Politics 45:39-43. Mayhew, David. 1974. Congress: The Electoral Connection. New Haven: Yale University Press. McCarty, Nolan, Keith T. Poole, and Howard Rosenthal. 2009. “Does Gerrymandering Cause Polarization?” American Journal of Political Science 53:666-80. McCarthy, Devin. 2019. “The Limits of Partisanship in Citizen Preferences on Redistricting.” <https://cpb-us-w2.wpmucdn.com/web.sas.upenn.edu/dist/7/538/files/2019 /07/Redistricting-Opinion_6.29.pdf>. McDonald, Michael P. 2004. “A Comparative Analysis of Redistricting Institutions in the United States, 2001-02.” State Politics and Policy Quarterly 4:371-95. McDonald, Michael. 2006. “Drawing the Line on District Competition.” PS: Political Science and Politics 39 (1): 91–94. McGhee, Eric, and Shor, Boris. “Has the Top Two Primary Elected More Moderates?” 15, no. 4 (December 2017): 1053–1066. McKee, Seth C., Jeremy M. Teigen, and Mathieu Turgeon. 2006. “The Partisan Impact of Congressional Redistricting: The Case of Texas, 2001-03.” Social Science Quarterly 87:308-17. McKee, Seth. 2013. “Political Conditions and the Electoral Effects of Redistricting.” American Politics Research 41(4): 623–650. Minta, Michael and Valeria Sinclair-Chapman. 2013. “Diversity in Political Institutions and Congressional Responsiveness to Minority Interests.” Political Research Quarterly 66:127-40. Murphy, Chad and Antoine Yoshinaka. 2009. “Are Mapmakers Able to Target and Protect Congressional Incumbents?” American Politics Research 37:955-82. NCSL. 2019a. “Redistricing and the Supreme Court: The Most Significant Cases.” National Conference of State Legislatures, last modified April 25, 2019. 54 NCSL. 2019b. “Redistricing Criteria.” National Conference of State Legislatures,last modified April 23, 2019. https://www.ncsl.org/research/redistricting/redistricting-criteria.aspx NCSL. "State Primary Election Types". National Conference of State Legislatures. http://www.ncsl.org/research/elections-and-campaigns/primary-types.aspx (accessed January 30, 2021). Nielson, Lindsay, and Visalvanich, Neil. “Primaries and Candidates: Examining the Influence of Primary Electorates on Candidate Ideology.” Political Science Research and Methods 5, no. 2 (April 1, 2017): 397–408. Norris, Pippa. 2014. Why Electoral Integrity Matters. New York: Cambridge University Press. Oppenheimer, D. M., Meyvis, T., & Davidenko, N. (2009). “Instructional manipulation checks: Detecting satisficing to increase statistical power.” Journal of Experimental Social Psychology, 45(4), 867–872. “Oppose Voter ID Legislation - Fact Sheet”. 2017. ACLU. Retrieved from https://www.aclu.org/other/oppose-voter-id-legislation-fact-sheet Panagopoulos, Costas. 2013. “Public Awareness and Attitudes about Redistricting.” Journal of Politics and Law 6:45-54. Peterson, Jordan Carr. 2019. “The Mask of Neutrality: Judicial Partisan Calculation and Legislative Redistricting.” Law & Policy 41:336-59. Rawls, John. 1971. A Theory of Justice. Cambridge: Harvard University Press. Riker, William H. 1980. “Implications from the Disequilibrium of Majority Rule for the Study of Institutions.” American Political Science Review 74:432-46. Rogowski, Jon C, and Langella, Stephanie. 2015. “Primary Systems and Candidate Ideology: Evidence From Federal and State Legislative Elections.” American Politics Research. Sadhwani, Sara, and Jane Junn. 2018. “Structuring Good Representation: Institutional Design and Elections in California.” PS: Political Science & Politics 51(2): 318–322. “State by State Redistricting Prcedures.” 2020. Ballotpedia. Retrieved from https://ballotpedia.org/State-by-state_redistricting_procedures Stephanopoulos, Nicholas O. and Eric McGhee. 2014. “Partisan Gerrymandering and the Efficiency Gap.” University of Chicago Law Review. Tate, Katherine. 2003. “Black Opinion on the Legitimacy of Racial Redistricting and Minority- Majority Districts.” American Political Science Review 97:45-56. Theriault, Sean. 2008. Party Polarization in Congress. New York: Cambridge University Press. 55 Thompson, Dennis F. 2002. Just Elections. Chicago: University of Chicago Press. Thomsen, Danielle M. 2020. “Parties and Polarization in Congress.” In Jamie L. Carson and Michael S. Lynch, eds., New Directions in Congressional Politics. Thurber, James A. and Antoine Yoshinaka. 2015. American Gridlock: The Sources, Character, and Impact of Political Polarization. New York: Cambridge University Press. Tolbert, Caroline J., Daniel A. Smith, and John C. Green 2009. “Strategic Voting and Legislative Redistricting Reform.” Political Research Quarterly 62:92-109. Tufte, Edward R. 1973. “The Relationship between Seats and Votes in Two-Party Systems.” American Political Science Review 67:540–54. Williamson, Ryan D. 2019. “Examining the Effects of Partisan Redistricting on Candidate Entry Decisions.” Election Law Journal 18:214-26. Winburn, Jonathan, Michael Henderson, and Conor M. Dowling. 2017. “From the Constituent’s Eye: Experimental Evidence on the District Selection Preferences of Individuals.” Political Research Quarterly 70:32-41. Yoshinaka, Antoine and Chad Murphy. 2009. “Partisan Gerrymandering and Population Instability: Completing the Redistricting Puzzle.” Political Geography 28:451-62. Yoshinaka, Antoine and Chad Murphy. 2011. “The Paradox of Redistricting: How Partisan Mapmakers Foster Competition.” Political Research Quarterly 64:435-47 56 Appendices Independent Redistricting Commissions are Associated with More Competitive Elections Appendix A: Additional regression models. Table A1: The regression models shown in Table 3 had a quality challenger control variable, southern state control variable, state and year fixed effects, and state clustered standard errors. The following models do not have a quality challenger control variable, southern state control variable, fixed effects, and do not recalculate standard errors. The results are consistent with the models found in the main text. Table A2: In Table 3 an election was considered competitive if it had a democratic vote share between 45 and 55 percent. The following models use alternate specifications for competitiveness. The results are consistent with the models found in the main text. 57 Table A3: Democratic vote share was considered 100 or 0 for uncontested and same-party elections in Table 3. The following model removes uncontested and same-party elections. The results are consistent with the models found in the main text. Table A4: The regression models shown in Table 3 included election data from 1982-2018. The following models only include election data from 2002-2018.The results are consistent with the models found in the main text. 58 Table A5: In the following model quality challenger is considered as a dependent variable. There is not a statistically significant relationship between redistricting institutions and quality challengers. Table A6: The following model has a binary dependent variable, Presidential Competitive, that identifies whether a district was competitive in the last presidential election (1 if Democrat’s share of two-party presidential vote was between 45% and 55%; 0 otherwise). There is not a statistically significant relationship between redistricting institutions and presidential competition. 59 Independent Redistricting Commissions Increase Voter Perceptions of Fairness Appendix B: Replicating results from Table 5 in text, but conditional on party identification of respondents, experiment 1. I present results of the experiment conditional on the political party of the respondent as some have argued that Republican voters may be less likely to support democracy reforms such as independent commissions. In addition, since the Democratic party controls the California state legislature, it is possible that there may be variation by party identification on the state legislative treatment. Generally, though, as can be seen below, the independent redistricting commission treatment leads to increased perceptions of fairness relative to the control condition across voters of all partisan identifications (Democrat, Republican and independent). Further, Democratic and independent respondents perceived the commission treatment as fairer than the state legislative treatment condition. Respondent party is measured by asking voters if they identify as Democrats, Republicans or independents/something else. Table B1: Independent Commissions Have Greater Perceptions of Fairness, Democratic respondents only Dep. var.: “How fair is California’s method of redrawing election district lines?” 1=Respondent says redistricting method is fair; 0=respondent says not fair Respondents randomly assigned and told… Percentage of respondents saying the way state draws electoral lines is fair Difference of means between treatment and control conditions Independent redistricting commission draws lines 83.4% 24.4** State legislature draws lines 44.0% -15.0** Control group (no information) 59.0% ---------- **p<0.01; *p<0.05. 60 Table B2: Independent Commissions Have Greater Perceptions of Fairness, Republican respondents only Dep. var.: “How fair is California’s method of redrawing election district lines?” 1=Respondent says redistricting method is fair; 0=respondent says not fair Respondents randomly assigned and told… Percentage of respondents saying the way state draws electoral lines is fair Difference of means between treatment and control conditions Independent redistricting commission draws lines 53.9% 34.5** State legislature draws lines 18.9% -0.5 Control group (no information) 19.4% ---------- **p<0.01; *p<0.05. Table B3: Independent Commissions Have Greater Perceptions of Fairness, independent respondents only Dep. var.: “How fair is California’s method of redrawing election district lines?” 1=Respondent says redistricting method is fair; 0=respondent says not fair Respondents randomly assigned and told… Percentage of respondents saying the way state draws electoral lines is fair Difference of means between treatment and control conditions Independent redistricting commission draws lines 70.2% 33.1** State legislature draws lines 22.4% -14.7** Control group (no information) 37.1% ---------- **p<0.01; *p<0.05. 61 Appendix C: Replicating results from Table 6 in text, but conditional on party identification of respondents, experiment 1. I present results of the experiment conditional on the political party of the respondent. Generally, as can be seen below, the independent redistricting commission treatment leads respondents to be less likely to think that the map drawing will unfairly favor one political party when compared to the control condition and the state legislature condition. Democratic, Republican, and independent voters all are less likely to think the lines will favor one party in the independent redistricting commission treatment. There is less evidence that voters respond consistently to the state legislature treatment relative to the control group. Respondent party is measured by asking voters if they identify as Democrats, Republicans or independents/something else. Table C1: Independent Commissions Reduce Voter Expectations that Electoral Lines Will be Unfairly Drawn to Favor One Political Party, Democratic respondents only Dep. var. “How likely do you think that California’s election districts will be unfairly drawn to favor one political party?” 7= extremely likely; 6=Somewhat likely; 5=Slightly likely; 4=Neither likely nor unlikely; 3=Slightly unlikely; 2=Somewhat unlikely; 1=Extremely unlikely Respondents randomly assigned and told… Mean belief that state will draw electoral lines to favor one party Difference of means between treatment and control conditions Independent redistricting commission draws lines 4.05 -0.48** State legislature draws lines 4.77 0.24* Control group (no information) 4.53 ---------- **p<0.01; *p<0.05. 62 Table C2: Independent Commissions Reduce Voter Expectations that Electoral Lines Will be Unfairly Drawn to Favor One Political Party, Republican respondents only Dep. var. “How likely do you think that California’s election districts will be unfairly drawn to favor one political party?” 7= extremely likely; 6=Somewhat likely; 5=Slightly likely; 4=Neither likely nor unlikely; 3=Slightly unlikely; 2=Somewhat unlikely; 1=Extremely unlikely Respondents randomly assigned and told… Mean belief that state will draw electoral lines to favor one party Difference of means between treatment and control conditions Independent redistricting commission draws lines 5.55 -0.58** State legislature draws lines 5.99 -0.14 Control group (no information) 6.13 ---------- **p<0.01; *p<0.05. Table C3: Independent Commissions Reduce Voter Expectations that Electoral Lines Will be Unfairly Drawn to Favor One Political Party, independent respondents only Dep. var. “How likely do you think that California’s election districts will be unfairly drawn to favor one political party?” 7= extremely likely; 6=Somewhat likely; 5=Slightly likely; 4=Neither likely nor unlikely; 3=Slightly unlikely; 2=Somewhat unlikely; 1=Extremely unlikely Respondents randomly assigned and told… Mean belief that state will draw electoral lines to favor one party Difference of means between treatment and control conditions Independent redistricting commission draws lines 4.53 -0.71** State legislature draws lines 5.33 0.09 Control group (no information) 5.24 ---------- **p<0.01; *p<0.05. 63 Appendix D: Replicating results from Table 7 in text, but conditional on party identification of respondents, experiment 1. I present results of the experiment conditional on the political party of the respondent. Generally, as can be seen below, the independent redistricting commission treatment leads respondents to be less likely to say that partisan gerrymandering is a problem in their state compared to the control condition and the state legislature condition, but the strongest effects are only with Democratic respondents. Democratic voters are less likely to say partisan gerrymandering is a problem in the independent redistricting commission treatment relative to the control condition. Interestingly, there were no statistically significant differences among Republican respondents across any of the randomized conditions. Among independent respondents, there were no statistically meaningful differences between treatment conditions and the control group among independent voters. Respondent party is measured by asking voters if they identify as Democrats, Republicans or independents/something else. Unlike all other outcome variables where Democratic, independent, and Republican respondents all responded similarly to the independent commission treatment relative to the control condition, in this survey outcome measure, Democratic respondents were more responsive to the independent commission treatment. Table D1: Independent Commissions Reduce Voter Perceptions that Partisan Gerrymandering is a Problem, Democratic respondents only Dep. var.: “Partisan gerrymandering is when elected legislators draw electoral lines in the legislature that favor their political party. Is partisan gerrymandering a problem in California?” 1= Definitely a problem; 2= Somewhat a problem; 3=Neither a problem nor not a problem; 4=Somewhat not a problem; 5=Definitely not a problem Respondents randomly assigned and told… Mean response indicating if partisan gerrymandering is problem Difference of means between treatment and control conditions Independent redistricting commission draws lines 2.95 0.27* State legislature draws lines 2.73 0.05 Control group (no information) 2.68 ---------- **p<0.01; *p<0.05. 64 Table D2: Independent Commissions Reduce Voter Perceptions that Partisan Gerrymandering is a Problem, Republican respondents only Dep. var.: “Partisan gerrymandering is when elected legislators draw electoral lines in the legislature that favor their political party. Is partisan gerrymandering a problem in California?” 1= Definitely a problem; 2= Somewhat a problem; 3=Neither a problem nor not a problem; 4=Somewhat not a problem; 5=Definitely not a problem Respondents randomly assigned and told… Mean response indicating if partisan gerrymandering is problem Difference of means between treatment and control conditions Independent redistricting commission draws lines 1.69 -0.03 State legislature draws lines 1.63 -0.09 Control group (no information) 1.72 ---------- **p<0.01; *p<0.05. Table D3: Independent Commissions Reduce Voter Perceptions that Partisan Gerrymandering is a Problem, independent respondents only Dep. var.: “Partisan gerrymandering is when elected legislators draw electoral lines in the legislature that favor their political party. Is partisan gerrymandering a problem in California?” 1= Definitely a problem; 2= Somewhat a problem; 3=Neither a problem nor not a problem; 4=Somewhat not a problem; 5=Definitely not a problem Respondents randomly assigned and told… Mean response indicating if partisan gerrymandering is problem Difference of means between treatment and control conditions Independent redistricting commission draws lines 2.55 0.14 State legislature draws lines 2.25 -0.16 Control group (no information) 2.41 ---------- **p<0.01; *p<0.05. 65 Appendix E: Results from Experiment 2 (Replication) Table E1: Experiment 2/Replication Results: Independent Commissions Have Greater Perceptions of Fairness Dep. var.: “How fair is California’s method of redrawing election district lines?” 1=Respondent says redistricting method is fair; 0=respondent says not fair Respondents randomly assigned and told… Percentage of respondents saying the way state draws electoral lines is fair Difference of means between treatment and control conditions Independent redistricting commission draws lines 75.6% 40.4** State legislature draws lines 25.8% -9.4** Control group (no information) 35.2% ---------- **p<0.01; *p<0.05. N=1016. Table E2:. Experiment 2/Replication Results: Independent Commissions Reduce Voter Expectations that Electoral Lines Will be Unfairly Drawn to Favor One Political Party Dep. var. “How likely do you think that California’s election districts will be unfairly drawn to favor one political party?” 7= extremely likely; 6=Somewhat likely; 5=Slightly likely; 4=Neither likely nor unlikely; 3=Slightly unlikely; 2=Somewhat unlikely; 1=Extremely unlikely Respondents randomly assigned and told… Mean belief that state will draw electoral lines to favor one party Difference of means between treatment and control conditions Independent redistricting commission draws lines 4.90 -.26** State legislature draws lines 5.55 0.39** Control group (no information) 5.16 ---------- **p<0.01; *p<0.05. N=1016. 66 Table E3: Experiment 2/Replication Results: Independent Commissions Reduce Voter Perceptions that Partisan Gerrymandering is a Problem Dep. var.: “Partisan gerrymandering is when elected legislators draw electoral lines in the legislature that favor their political party. Is partisan gerrymandering a problem in California?” 1= Definitely a problem; 2= Somewhat a problem; 3=Neither a problem nor not a problem; 4=Somewhat not a problem; 5=Definitely not a problem Respondents randomly assigned and told… Mean response indicating if partisan gerrymandering is problem Difference of means between treatment and control conditions Independent redistricting commission draws lines 2.60 0.05 State legislature draws lines 2.27 -0.28** Control group (no information) 2.55 ---------- **p<0.01; *p<0.05. N=1016. 67 Appendix F: Replicating results from Table 5 in text, but conditional on party identification of respondents, experiment 2/replication (first outcome variable). Table F1: Experiment 2/Independent Commissions Have Greater Perceptions of Fairness, Democratic respondents only Dep. var.: “How fair is California’s method of redrawing election district lines?” 1=Respondent says redistricting method is fair; 0=respondent says not fair Respondents randomly assigned and told… Percentage of respondents saying the way state draws electoral lines is fair Difference of means between treatment and control conditions Independent redistricting commission draws lines 86.1% 42.8** State legislature draws lines 30.7% -12.6** Control group (no information) 43.3% ---------- **p<0.01; *p<0.05. Table F2: Experiment 2/Independent Commissions Have Greater Perceptions of Fairness, Republican respondents only Dep. var.: “How fair is California’s method of redrawing election district lines?” 1=Respondent says redistricting method is fair; 0=respondent says not fair Respondents randomly assigned and told… Percentage of respondents saying the way state draws electoral lines is fair Difference of means between treatment and control conditions Independent redistricting commission draws lines 71.8% 38.0** State legislature draws lines 29.2% -4.6 Control group (no information) 33.8% ---------- **p<0.01; *p<0.05. 68 Table F3: Experiment 2/Independent Commissions Have Greater Perceptions of Fairness, independent respondents only Dep. var.: “How fair is California’s method of redrawing election district lines?” 1=Respondent says redistricting method is fair; 0=respondent says not fair Respondents randomly assigned and told… Percentage of respondents saying the way state draws electoral lines is fair Difference of means between treatment and control conditions Independent redistricting commission draws lines 57.5% 33.0** State legislature draws lines 16.8% -7.7* Control group (no information) 24.5% ---------- **p<0.01; *p<0.05. 69 Appendix G: Replicating results from Table 6 in text, but conditional on party identification of respondents, experiment 2/replication (second outcome variable) Table G1: Experiment 2/Independent Commissions Reduce Voter Expectations that Electoral Lines Will be Unfairly Drawn to Favor One Political Party, Democratic respondents only Dep. var. “How likely do you think that California’s election districts will be unfairly drawn to favor one political party?” 7= extremely likely; 6=Somewhat likely; 5=Slightly likely; 4=Neither likely nor unlikely; 3=Slightly unlikely; 2=Somewhat unlikely; 1=Extremely unlikely Respondents randomly assigned and told… Mean belief that state will draw electoral lines to favor one party Difference of means between treatment and control conditions Independent redistricting commission draws lines 4.66 -0.25* State legislature draws lines 5.41 0.50** Control group (no information) 4.91 ---------- **p<0.01; *p<0.05. Table G2: Experiment 2/Independent Commissions Reduce Voter Expectations that Electoral Lines Will be Unfairly Drawn to Favor One Political Party, Republican respondents only Dep. var. “How likely do you think that California’s election districts will be unfairly drawn to favor one political party?” 7= extremely likely; 6=Somewhat likely; 5=Slightly likely; 4=Neither likely nor unlikely; 3=Slightly unlikely; 2=Somewhat unlikely; 1=Extremely unlikely Respondents randomly assigned and told… Mean belief that state will draw electoral lines to favor one party Difference of means between treatment and control conditions Independent redistricting commission draws lines 5.22 -0.15 State legislature draws lines 5.69 0.32* Control group (no information) 5.37 ---------- **p<0.01; *p<0.05. 70 Table G3: Experiment 2/Independent Commissions Reduce Voter Expectations that Electoral Lines Will be Unfairly Drawn to Favor One Political Party, independent respondents only Dep. var. “How likely do you think that California’s election districts will be unfairly drawn to favor one political party?” 7= extremely likely; 6=Somewhat likely; 5=Slightly likely; 4=Neither likely nor unlikely; 3=Slightly unlikely; 2=Somewhat unlikely; 1=Extremely unlikely Respondents randomly assigned and told… Mean belief that state will draw electoral lines to favor one party Difference of means between treatment and control conditions Independent redistricting commission draws lines 5.14 -0.25* State legislature draws lines 5.65 0.26* Control group (no information) 5.39 ---------- **p<0.01; *p<0.05. 71 Appendix H: Replicating results from Table 7 in text, but conditional on party identification of respondents, experiment 2/replication (third outcome variable). Table H1: Experiment 2/Independent Commissions Reduce Voter Perceptions that Partisan Gerrymandering is a Problem, Democratic respondents only Dep. var.: “Partisan gerrymandering is when elected legislators draw electoral lines in the legislature that favor their political party. Is partisan gerrymandering a problem in California?” 1= Definitely a problem; 2= Somewhat a problem; 3=Neither a problem nor not a problem; 4=Somewhat not a problem; 5=Definitely not a problem Respondents randomly assigned and told… Mean response indicating if partisan gerrymandering is problem Difference of means between treatment and control conditions Independent redistricting commission draws lines 2.88 0.02 State legislature draws lines 2.51 -0.35** Control group (no information) 2.86 ---------- **p<0.01; *p<0.05. Table H2: Experiment 2/Independent Commissions Reduce Voter Perceptions that Partisan Gerrymandering is a Problem, Republican respondents only Dep. var.: “Partisan gerrymandering is when elected legislators draw electoral lines in the legislature that favor their political party. Is partisan gerrymandering a problem in California?” 1= Definitely a problem; 2= Somewhat a problem; 3=Neither a problem nor not a problem; 4=Somewhat not a problem; 5=Definitely not a problem Respondents randomly assigned and told… Mean response indicating if partisan gerrymandering is problem Difference of means between treatment and control conditions Independent redistricting commission draws lines 1.97 -0.25 State legislature draws lines 1.89 -0.33* Control group (no information) 2.22 ---------- **p<0.01; *p<0.05. 72 Table H3: Experiment 2/Independent Commissions Reduce Voter Perceptions that Partisan Gerrymandering is a Problem, independent respondents only Dep. var.: “Partisan gerrymandering is when elected legislators draw electoral lines in the legislature that favor their political party. Is partisan gerrymandering a problem in California?” 1= Definitely a problem; 2= Somewhat a problem; 3=Neither a problem nor not a problem; 4=Somewhat not a problem; 5=Definitely not a problem Respondents randomly assigned and told… Mean response indicating if partisan gerrymandering is problem Difference of means between treatment and control conditions Independent redistricting commission draws lines 2.51 0.20* State legislature draws lines 2.16 -0.15 Control group (no information) 2.31 ---------- **p<0.01; *p<0.05. 73 Top-Two Primaries Can Moderate Election Outcomes Appendix I: Additional Demographic Information Table I1: Ethnicity There was balance between treatment conditions for ethnicity. Respondents with multiple ethnicities are counted multiple times. Black Hispanic Asian White Other Top-Two 11% 8% 6% 72% 3% Closed 12% 9% 6% 70% 3% Table I2: Education There was balance between treatment conditions for education. Top-Two Closed Associate degree in college 13% 13% Bachelor’s degree in college 23% 25% Doctoral degree 1% 2% High school graduate 28% 26% Less than high school degree 3% 2% Master’s degree 9% 11% Professional degree (JD,MD) 1% 1% Some college but no degree 23% 21% 74 Appendix J: General Election Results by Treatment Condition Table J1: Candidate A vs Candidate B: Which candidate did respondents prefer? Candidate A (Republican) Candidate B (Republican) Center-Right Far-Right Top-Two (All voters) 71% 29% Closed (All voters) 69% 31% Top-Two (Republicans) 66% 34% Closed (Republicans) 63% 37% Top-Two (Democrats) 75% 25% Closed (Democrats) 72% 28% Table J2: Candidate C vs Candidate D: Which candidate did respondents prefer? Candidate C (Democrat) Candidate D (Democrat) Far-Left Center-Left Top-Two (All voters) 39% 61% Closed (All voters) 37% 63% Top-Two (Republicans) 28% 72% Closed (Republicans) 26% 74% Top-Two (Democrats) 47% 53% Closed (Democrats) 48% 52% 75 Table J3: Candidate A vs Candidate C: Which candidate did respondents prefer? Candidate A (Republican) Candidate C (Democrat) Center-Right Far-Left Top-Two (All voters) 47% 53% Closed (All voters) 48% 52% Top-Two (Republicans) 81% 19% Closed (Republicans) 79% 21% Top-Two (Democrats) 23% 77% Closed (Democrats) 28% 72% Table J4: Candidate A vs Candidate D: Which candidate did respondents prefer? Candidate A (Republican) Candidate D (Democrat) Center-Right Center-Left Top-Two (All voters) 43% 57% Closed (All voters) 44% 56% Top-Two (Republicans) 75% 25% Closed (Republicans) 74% 26% Top-Two (Democrats) 22% 78% Closed (Democrats) 26% 74% 76 Table J5: Candidate B vs Candidate C: Which candidate did respondents prefer? Candidate B (Republican) Candidate C (Democrat) Far-Right Far-Left Top-Two (All voters) 41% 59% Closed (All voters) 43% 57% Top-Two (Republicans) 73% 27% Closed (Republicans) 72% 28% Top-Two (Democrats) 19% 81% Closed (Democrats) 25% 75% Table J6: Candidate B vs Candidate D: Which candidate did respondents prefer? Candidate B (Republican) Candidate D (Democrat) Far-Right Center-Left Top-Two (All voters) 37% 63% Closed (All voters) 39% 61% Top-Two (Republicans) 67% 33% Closed (Republicans) 66% 34% Top-Two (Democrats) 18% 82% Closed (Democrats) 25% 75%

Abstract (if available)

Abstract Independent Redistricting Commissions are Associated with More Competitive Elections.
Election competition is crucial for democratic accountability, yet most U.S. House elections are uncompetitive. Using district-level data from 1982 to 2018, I examine the relationship between redistricting institutions and U.S. House election competition. I separate independent and political commissions and find that relative to legislative redistricting, independent commissions are 1.98 times as likely to have competitive elections; and they decrease incumbent party wins by 60%.
Independent Redistricting Commissions Increase Voter Perceptions of Fairness
Do independent redistricting commissions cause voters to view the redistricting process as fair? Scholars have focused on the role that redistricting plays in political outcomes, but the process built into institutions also signals important values such as fairness. I argue that independent redistricting commissions improve voter attitudes toward fairness in the redistricting process. In two survey experiments, I find that voters are more likely to rate the redistricting process as fair and to believe that partisan gerrymandering is not a problem when they learn about their state’s independent redistricting commission. In contrast, voters in a control group and in a randomized treatment group who learn that state legislators conduct redistricting are less likely to view the redistricting process as fair.
Top-Two Primaries Can Moderate Election Outcomes
How does the design of primary systems affect voter decision-making and election outcomes? I conduct a survey experiment with participants randomly assigned to vote in either a Top-Two or Closed primary election and in a series of general election scenarios. I find that same-party general elections lead to potential voting coalitions between moderate members of one party and members of the opposite party. In this scenario of same-party general elections, Top-Two primary systems lead voters to choose more moderate candidates.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Politics is something we do together: identity and institutions in U.S. elections

PDF

Bureaucratic politics and power building in the administrative state

PDF

La eleccion de la pandemia: analyzing Latino political behavior during the 2020 election

PDF

Descriptive representation in a multi-racial America

PDF

Private interests in American government institutions

PDF

Are you as old as you feel? A quantitative and qualitative analysis of leader age and foreign policy decision-making

PDF

Essays on political economy and corruption

PDF

Voices unheard, stories untold: Black women, police violence and political participation

PDF

Local votes and outside money: campaign contribution geographic origins and their impact on Los Angeles City Council election outcomes

PDF

Voter engagement in the 2020 presidential election: what we can learn from the voter engagement initiatives in California-based and national nonpartisan organizations to increase voter participation

PDF

The small dollar political donor: why regular folks give money to politics

PDF

By word, not deed: advancing a theory of rhetorical compliance in international law

PDF

The representation of LGBT interests in the United States: a multimethod approach

PDF

The strategy of firms towards political elites and consumers in polarized western democracies

PDF

In/visible constituents: the representation of undocumented immigrants

PDF

Three essays on the political economy of federal spending

PDF

Running for office? Three different models in U.S. political campaigns & elections

PDF

Pernicious racism and the impacts of intentional racist in a color-blind society

PDF

The IO contestation project: a multi-method approach towards understanding the ways states contest international organizations.

PDF

Signaling identity: how race and gender shape what representatives say online

Asset Metadata

Creator Nelson, Matthew (author)

Core Title Can election reforms improve democracy?

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Political Science and International Relations

Degree Conferral Date 2022-08

Publication Date 05/12/2022

Defense Date 04/20/2022

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag Elections,OAI-PMH Harvest,primary reform,redistricting,representation,voter behavior

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Grose, Christian (committee chair), Lo, James (committee member), McCann, Pamela (committee member)

Creator Email mdn817@yahoo.com,nels171@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC111313291

Unique identifier UC111313291

Document Type Dissertation

Format application/pdf (imt)

Rights Nelson, Matthew

Type texts

Source 20220517-usctheses-batch-942 (batch), University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright. The original signature page accompanying the original submission of the work to the USC Libraries is retained by the USC Libraries and a copy of it may be obtained by authorized requesters contacting the repository e-mail address given.

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA

Repository Email cisadmin@lib.usc.edu