Introduction
How does the mass public form attitudes on electoral rules and reforms? This question looms large in the literature on electoral systems, as rules and reforms shape election outcomes and provide a foundation for democratic governance. Voters often play a direct role in shaping policies on electoral reforms, such as legislative redistricting. In 2018, Michigan voters passed a statewide ballot initiative that transferred redistricting power from the state legislature to an independent redistricting commission. In 2024, a similar ballot measure failed in Ohio; voters elected to retain a partisan redistricting commission process over a proposed independent redistricting commission process. More broadly, public attitudes on electoral rules and reforms set the stage for whether self-interested party elites are able to amend election rules for their benefit (McCarthy Reference McCarthy2023). As such, public opinion on electoral reform represents a crucial quantity of interest for research on electoral systems and policymaking.
Public opinion research on attitudes toward electoral reform reveals a nuanced trade-off between principles and partisan self-interest. A long line of research, both observational and experimental, shows that partisan self-interest, the motivation to increase copartisans’ chances of electoral success, acts as a central driver of mass opinion on electoral reforms (Alarian and Zonszein Reference Alarian and Zonszein2024; Biggers Reference Biggers2019; Biggers and Bowler Reference Biggers and Bowler2022; Bowler and Donovan Reference Bowler and Donovan2016; Kane Reference Kane2017). However, recent studies show that public opinion on electoral reform is also shaped by other factors, such as the principle of fairness (Biggers and Bowler Reference Biggers and Bowler2023; McCarthy Reference McCarthy2022; Virgin Reference Virgin2023). This study leverages two preregistered survey experiments to explore how voters weigh principles against partisan self-interest when forming opinions on electoral reforms. Specifically, I clarify two aspects of the partisan self-interest motivation – one based on the nature of political representation, and one based on the concept of loss aversion.
First, I ask whether the public’s partisan self-interest motivation on electoral reform stems more from individual representation considerations or broader partisan power considerations. That is, do voters care more about the partisan affiliation of their individual representatives (dyadic representation) or the partisan balance of power in policymaking institutions (collective representation)? Drawing on research that highlights the high demand for collective representation (Harden and Clark Reference Harden and Clark2016) and the trends of increasing partisan polarization and the nationalization of American politics (Abramowitz and Webster Reference Abramowitz and Webster2016; Carson, Sievert and Williamson Reference Carson, Sievert and Williamson2023; Hopkins Reference Hopkins2018), I argue that partisan power considerations are a more potent force than those related to individual representation.
I use a survey experiment on state legislative redistricting to shed light on this question. Because redistricting has the potential to shift both individual and collective partisan power, it represents an electoral reform uniquely situated to examine how collective and dyadic representation considerations shape support for election reforms. Respondents were introduced to a newly proposed district map, explicitly framed as an improvement from a democratic principles standpoint, drawn by a non-partisan, non-profit organization. I implemented a 2 × 2 design, varying whether two forms of partisan self-interest – district-level individual representation and statewide partisan power – were triggered as an expected downside of the proposed map. These treatments involved informing respondents that the proposed map would likely lead to negative outcomes for the respondent’s party in future elections. Specifically, experimental treatments included information that the proposed map would enhance the out-party’s chances of winning in the respondent’s legislative district (individual representation treatment) and gaining seats in the state legislature (partisan power treatment). Respondents then chose between implementing the proposed district map and keeping their state’s current district map.
I find that both considerations substantially decrease public support for the proposed map, and the treatment effects are similar in magnitude. Over 50% of respondents who read only about the proposed map’s democratic virtues chose the proposed map over the status quo, while under 25% of respondents assigned to either partisan consideration treatment chose the proposed map. Therefore, I report little support for the preregistered prediction that partisan power considerations are a more powerful determinant of opposition to redistricting improvements than legislative representation considerations. I conclude that both partisan power and individual representation are powerful enough considerations to activate partisan self-interest in the formation of attitudes on electoral reform.
Second, I explore the principles versus partisan self-interest trade-off through the lens of loss aversion, a core component of prospect theory in behavioral economics, which dictates that prospective losses weigh more heavily on decision-making than prospective gains (Kahneman and Tversky Reference Kahneman and Tversky1979; Tversky and Kahneman Reference Tversky and Kahneman1981). I argue that the principle of loss aversion shapes public opinion on electoral reform. Electoral reforms lead to winners and losers, and I hypothesize that preventing partisan loss is a stronger incentive for voters’ to sacrifice democratic principles than achieving partisan gain.
Building on the state legislative redistricting design from the first experiment, I field an additional survey experiment to examine the loss aversion expectation. This experimental design includes three treatment groups: 1) Neutral – a nonpartisan map that better represents voters in the respondents’ state with no reference to electoral consequences; 2) Partisan Loss – a nonpartisan map that better represents voters in the respondents’ state but will likely hurt the respondents’ party in upcoming elections; 3) Partisan Gain – a partisan map that was criticized as unfair but will likely help the respondents’ party in upcoming elections. Results from this experiment suggest that preventing loss by resisting a new district map provides a more powerful incentive for Americans’ to violate democratic principles than achieving partisan gain via a newly drawn partisan map. Respondents tended to reject fair redistricting reform that would lead to partisan loss, but they also rejected unfair redistricting reform that would lead to partisan gain. The causal mechanism behind this finding remains unclear and requires further research, as both loss aversion and status quo bias offer viable explanations, but this study reveals a stark asymmetry between preventing partisan loss and achieving partisan gain as triggers of democratic principle violation.
Public support for electoral reform
Research on electoral rules and reform often focuses on political parties as the critical actors and partisan self-interest as the central determinant of public opinion. As the key players in the zero-sum game of electoral politics, parties establish policymaking power through electoral success. Electoral rules and processes influence parties’ electoral prospects, and parties thus stand to gain or lose an electoral advantage through electoral reforms. Therefore, rational self-interest shapes party elites’ support for or opposition to electoral rules and reforms (Benoit Reference Benoit2007; Boix Reference Boix1999; Bowler, Donovan and Karp Reference Bowler, Donovan and Karp2006; Hicks et al. Reference Hicks, McKee, Sellers and Smith2015; Riker Reference Riker1988). Recent research on electoral reform has shifted the focus from party elites to the mass public by exploring the determinants of public opinion on electoral reforms. This stream of research generally frames public opinion on electoral reforms as a tug-of-war between principles and partisan self-interest.
Most voters have partisan leanings and plausibly follow a similar framework as party elites when it comes to electoral reform. Party identification works as a crucial political identity and information shortcut for voters (Achen and Bartels Reference Achen and Bartels2017; Campbell et al. Reference Campbell, Converse, Miller and Stokes1960; Green, Palmquist and Schickler Reference Green, Palmquist and Schickler2002; Lau and Redlawsk Reference Lau and Redlawsk2006), leading public opinion to mirror party leaders (Zaller Reference Zaller1992). Elite signals link elite discourse on election policies and citizen preferences, as voters tend to lack an awareness of how electoral policies shape election results absent elite signaling (Bowler and Donovan Reference Bowler and Donovan2016; Coll Reference Coll2024; Gronke et al. Reference Gronke, Hicks, McKee, Stewart and Dunham2019). Further, American politics has become more polarized (Abramowitz and Saunders Reference Abramowitz and Saunders2008; Webster and Abramowitz Reference Webster and Abramowitz2017), hostile (Abramowitz and Webster Reference Abramowitz and Webster2016; Iyengar et al. Reference Iyengar, Lelkes, Levendusky, Malhotra and Westwood2019; Iyengar and Westwood Reference Iyengar and Westwood2015; Kalmoe and Mason Reference Kalmoe and Mason2022; Webster Reference Webster2020), and nationalized (Butler and Sutherland Reference Butler and Sutherland2023; Carson, Sievert and Williamson Reference Carson, Sievert and Williamson2023; Hopkins Reference Hopkins2018). Such developments heighten the likelihood that the public will view electoral reforms through the prism of partisan self-interest.
Indeed, both observational and experimental studies find partisan self-interest to play a major role in shaping public opinion on electoral reform. Observational studies reveal a robust association between partisan self-interest and public opinion on various electoral reforms, such as voter identification laws (Bowler and Donovan Reference Bowler and Donovan2016; Wilson and Brewer Reference Wilson and Brewer2013), electoral college reform (Aldrich, Reifler and Munger Reference Aldrich, Reifler and Munger2014; Alvarez et al. Reference Alvarez, Hall, Levin and Stewart2011), and making Election Day a holiday (Alvarez et al. Reference Alvarez, Hall, Levin and Stewart2011). Voters tend to support reforms that are likely to help their party’s electoral interests and oppose reforms that are likely to hurt their party’s electoral interests. On redistricting, Tolbert, Smith and Green (Reference Tolbert, Smith and Green2009) find voters who lose under the current institutional arrangement are more supportive of independent redistricting commissions, whereas voters who win under the current institutional arrangement oppose such reforms.
Recent experimental research identifies a causal effect of partisan self-interest on support for electoral reforms (Alarian and Zonszein Reference Alarian and Zonszein2024; Biggers Reference Biggers2019; Biggers and Bowler Reference Biggers and Bowler2022; Kane Reference Kane2017; McCarthy Reference McCarthy2019; Winburn, Henderson and Dowling Reference Winburn, Henderson and Dowling2017). Biggers and Bowler (Reference Biggers and Bowler2022) use panel experiments to show that partisans are more willing to support reforms they previously deemed unfair when such reforms help rather than hurt their party’s electoral prospects. Winburn et al. (Reference Winburn, Henderson and Dowling2017) come to a similar conclusion based on experimental designs pitting redistricting principles against partisan self-interest: “The results of two experiments indicate that use of county integrity as a criterion for selecting a preferred district configuration fades when people have access to partisan information, even when they are explicitly reminded to preserve geographical boundaries to the extent possible” (33). Alarian and Zonszein (Reference Alarian and Zonszein2024) show that public support for extending voting rights to non-citizens in local elections is conditional on partisan alignment between citizens and potential non-citizen voters. Clearly, voters consider electoral consequences and partisan self-interest when forming an opinion on electoral reforms.
However, factors beyond partisan self-interest also influence public opinion on electoral reforms. For instance, Virgin (Reference Virgin2023) shows that core values – egalitarianism, moral tolerance, self-reliance, and economic individualism – play a critical role in the determination of public opinion on electoral rules and reforms. Using an experimental design that varies both partisan considerations and core values considerations, Virgin finds evidence in favor of core values: “not only do they have an important effect net of partisan concerns, but also, they attenuate the effect of partisan self-interest in instances for which the two predispositions have been made to countervail” (253). Additionally, the mass public prefers procedural fairness when considering electoral rules and reforms. Biggers and Bowler (Reference Biggers and Bowler2023) show that priming fairness considerations decreases the influence of partisan self-interest. McCarthy (Reference McCarthy2022) leverages a novel experimental design on congressional redistricting – respondents were asked to choose between various district maps – and finds that procedural fairness often outweighs partisan self-interest when it comes to public opinion on redistricting. Partisanship matters, but public opinion on electoral reforms is more complex than partisan self-interest.
The literature on public opinion towards electoral reform suggests both principles and partisan self-interest influence attitudes, leading to questions on how the public goes about weighing this trade-off. Recent studies in this stream of research progress from exploring whether principles and partisan self-interest influence attitudes to exploring the mechanisms behind the public’s willingness to violate principles and democratic norms in the name of partisan self-interest. For instance, Biggers (Reference Biggers2019) shows that the mass public does not merely take cues from party elites to arrive at partisan self-interest on electoral reform opinions. Rather, voters independently take electoral outcomes into account while forming opinions on electoral reform. McCarthy (Reference McCarthy2023) explores a similar question on elite influence, pitting an “elite influence” model and a “principled citizen” model against each other across three experiments. McCarthy finds that the public tends to reject party elites’ explicit attempts at electoral manipulation, but elites can effectively invoke democratic principles to justify the strategic electoral policy to copartisans in the electorate. In this paper, I advance this stream of research by investigating the conditions under which the public becomes more likely to sacrifice principles in the name of partisan self-interest.
In addition to the topical focus of electoral reform, this stream of research advances a broader research agenda on the relationship between partisanship, political values, and democratic norms. The trade-off between principles and partisan self-interest around electoral reform highlights fundamental questions about American democracy. Recent studies grapple with the interaction of partisanship, values, and democratic norms (Alarian and Zonszein Reference Alarian and Zonszein2024; Druckman et al. Reference Druckman, Kang, Chu, Stagnaro, Voelkel, Mernyk, Pink, Redekopp, Rand and Willer2023; Goren Reference Goren2005; Goren, Federico and Kittilson Reference Goren, Federico and Kittilson2009; Graham and Svolik Reference Graham and Svolik2020; Grillo and Prato Reference Grillo and Prato2023; Holliday et al. Reference Holliday, Iyengar, Lelkes and Westwood2024; Levitsky and Ziblatt Reference Levitsky and Ziblatt2018), and research on electoral reform contributes critical insights on this topic (e.g. Biggers and Bowler Reference Biggers and Bowler2022; Virgin Reference Virgin2023). I build on this foundation of research, using electoral reform as a topical focus to explore the extent to which partisanship overrides American considerations of democratic norms when it comes to structuring key democratic institutions.
Theory
The literature discussed above suggests a trade-off between principles and partisan self-interest shapes public opinion on electoral reforms. Of course, the nature of this trade-off differs based on the electoral reform, the political environment, and question framing. Further research is needed to clarify how voters weigh partisan self-interest and principles when forming opinions on electoral reforms. I explore two aspects of the partisanship versus principles trade-off in this study – one based on the nature of political representation, and one based on the principle of loss aversion.
My first research question aims to shed additional light on the partisan self-interest motivation behind public opinion on electoral reform: are individual representation considerations or partisan power considerations more powerful in determining voters’ stance on electoral reform? In other words, does voters’ partisan self-interest stem more from individual representation – whether they are represented by a copartisan – or broader party strength – whether their party is gaining power? On one hand, legislative representation is a core component of American democracy, and voters’ representatives are their voice in government. Therefore, representation by a copartisan in the US Congress or their state’s legislature should be central to voters’ consideration of electoral reforms. On the other hand, the nationalization of American politics centers partisan power dynamics, rather than individual representation, in the minds of voters. As such, voters likely care more about partisan power than individual representation.
Both considerations plausibly factor into public opinion on electoral reform. However, I theorize that voters care more about partisan power dynamics than individual representation. This claim is grounded in two relevant streams of research. First, studies on legislative representation highlight two pathways to representation: dyadic representation and collective representation. Dyadic representation focuses on constituents’ relationship with the individual legislator representing their electoral district (Miller and Stokes Reference Miller and Stokes1963), while collective representation focuses on constituents’ relationship with a group of legislators with whom they share a common trait or belief (Weissberg Reference Weissberg1978). Individual representation considerations map onto the concept of dyadic representation, while partisan power considerations map onto the concept of collective representation. Harden and Clark (Reference Harden and Clark2016) leverage survey experiments on both race and partisanship to show that there is greater public demand for collective representation than dyadic representation; collective representation is more likely to enhance feelings of representation. Based on this research, I expect partisan power considerations to carry greater weight than individual representation considerations in the formation of public opinion on electoral reform.
Second, the politics of previous eras may have led people to focus more on individual representation (Fenno Reference Fenno1978), but polarization and nationalization have transformed American politics to prioritize partisan allegiance (Abramowitz and Webster Reference Abramowitz and Webster2016; Hopkins Reference Hopkins2018; Mason Reference Mason2018). As such, I argue voters interpret electoral reform proposals through a prism of partisan power rather than individual representation.
Hypothesis 1: Partisan power considerations are a more powerful determinant of opposition to redistricting improvements than legislative representation considerations.
My second research question further probes the mechanism behind the public’s willingness to violate democratic principles in the name of partisanship: is partisan gain or partisan loss a more powerful incentive for violating democratic principles? In other words, are voters more willing to sacrifice principles in order to prevent their party from losing power or to help their party gain power? The literature on behavioral economics and prospect theory, a model of human decision-making under risk and uncertainty (Kahneman and Tversky Reference Kahneman and Tversky1979; Tversky and Kahneman Reference Tversky and Kahneman1991), offers useful insights into this question. I specifically focus on the concept of loss aversion, the idea that prospective losses weigh more heavily on decision-making than prospective gains. Tversky and Kahneman (Reference Tversky and Kahneman1981) show that people are risk-avoidant when facing prospective gains and risk-seeking when facing prospective losses. The principle of loss aversion has been studied and confirmed in a wide variety of decision-making settings (Ericson and Fuster Reference Ericson and Fuster2014; Gächter, Johnson and Herrmann Reference Gächter, Johnson and Herrmann2022; Kahneman, Knetsch and Thaler Reference Kahneman, Knetsch and Thaler1990; Wakker Reference Wakker2010).
Loss aversion, as well as other elements of prospect theory, have been effectively incorporated into political science research (Druckman and Lupia Reference Druckman and Lupia2000; McDermott Reference McDermott2004). On voting and elections, loss aversion has been shown to lead to an incumbency advantage (Quattrone and Tversky Reference Quattrone and Tversky1988), the president’s party’s tendency to lose seats in midterm elections (Patty Reference Patty2006), campaign platform rigidity (Lockwood and Rockey Reference Lockwood and Rockey2020), and a status quo bias in constitutional design (Attanasi, Corazzini and Passarelli Reference Attanasi, Corazzini and Passarelli2017). Alesina and Passarelli (Reference Alesina and Passarelli2019) develop spatial voting models that combine policy preferences and loss aversion to generate insights into political outcomes, such as the differential power of status quo bias on young and old societies. Additionally, a stream of research on framing finds that loss aversion plays a key role in how the public forms opinions on public policy. Voters are more supportive of policies when they are framed around preventing loss than securing gains, and this finding holds across numerous policies and research designs (Arceneaux Reference Arceneaux2012; Druckman and McDermott Reference Druckman and McDermott2008; Jerit Reference Jerit2009; Kam and Simas Reference Kam and Simas2010; Lau Reference Lau1985; Osmundsen and Petersen Reference Osmundsen and Petersen2020). In sum, loss aversion looms large in political decision-making and opinion formation.
I argue that the principle of loss aversion – that people are risk avoidant when facing prospective gains and risk-seeking when facing prospective losses – shapes opinion formation on electoral reform. In this case, risk is conceptualized as the deterioration of democratic norms and institutions, which could destabilize American democracy. Electoral reform leads to winners and losers, and the threat of partisan loss is likely a stronger incentive for voters to sacrifice democratic principles than the opportunity for partisan gain. In addition to the established literature on loss aversion, a robust stream of research on affective polarization shows that negative partisanship accounts for the recent increase in the distance between partisans (Abramowitz and Webster Reference Abramowitz and Webster2016; Iyengar, Sood and Lelkes Reference Iyengar, Sood and Lelkes2012; Iyengar and Westwood Reference Iyengar and Westwood2015). Americans increasingly display animus towards the opposing party (Iyengar et al. Reference Iyengar, Lelkes, Levendusky, Malhotra and Westwood2019), and this dynamic likely enhances the effects of loss aversion when it comes to public opinion on electoral reforms. Considering the twin pressures of loss aversion and negative partisanship, I argue that Americans prioritize preventing the opposing party’s electoral success over enhancing their own party’s electoral success. As such, people are more likely to choose partisan self-interest over democratic principles when facing potential loss, as compared to potential gain, relative to the status quo.
Hypothesis 2: Preventing partisan loss is a more powerful driver of Americans’ willingness to violate democratic principles than achieving partisan gain.
Combined, my two hypotheses offer testable implications of an overarching theory on the mechanisms behind partisan self-interest and public opinion on electoral reform. In an era of nationalized politics and negative partisanship, I argue that partisan self-interest is motivated by broad party power over individual representation (H1) and fear of partisan loss over-enthusiasm for partisan gain (H2). I leverage two survey experiments to test these hypotheses, thereby generating insights into the nature of partisan self-interest as a determinant of public opinion on electoral reform.
Experiment #1: Clarifying the concept of partisan self-interest
My first research question asks whether individual representation considerations or partisan power considerations serve as a more powerful driver of public opinion on electoral reform. I argue that partisan power considerations are more likely to trigger partisan self-interest, and I test this proposition through a survey experiment on state legislative redistricting reform. Redistricting is a critical element of American elections, periodically reshaping the electoral landscape of state and national legislatures and significantly influencing the nature of legislative representation (Kaslovsky and Kistner Reference Kaslovsky and Kistner2024; Keena et al. Reference Keena, Latner, McGann and Smith2021). Further, electoral reform movements have recently targeted legislative redistricting. Reformers have drawn attention to the practice of gerrymandering, where politicians draw legislative district maps with the goal of partisan self-interest, and recent reforms have succeeded in transferring the power to draw legislative districts from state legislators to independent redistricting committees (Gartner Reference Gartner2019; Timm Reference Timm2021). Redistricting offers an ideal platform to test my hypothesis, as it presents an opportunity to realistically trigger both individual representation and partisan power considerations. Additionally, public awareness of redistricting is generally low (Panagopoulos Reference Panagopoulos2013), which allows for believable experimental treatment vignettes.
Experimental design
I use an online survey experiment to explore whether legislative representation considerations or broader partisan power considerations are a more powerful driver of opposition to a positive reform, from a democratic principles perspective, to state legislative redistricting reform.Footnote 1 Roughly 1,500 respondents were recruited through the Lucid Theorem survey platform in October of 2023 to participate in the experiment. Lucid samples are not randomly drawn, but they provide comparable demographic balance and experimental estimates to nationally representative samples (Coppock and McClellan Reference Coppock and McClellan2019).Footnote 2
After answering an assortment of demographic questions, survey respondents were asked to read an experimental vignette on state legislative redistricting in their home state. The vignette introduced respondents to a newly proposed state legislative district map and explained that the proposed map, which was drawn by “a strictly non-partisan, non-profit organization,” more accurately reflects voters in their state than the current district map. Thus, the proposed map was explicitly framed for all respondents as an improvement over the status quo from a democratic principles standpoint. In other words, the democratic principles signal is constant across all treatment conditions and points towards the proposed map.
Next, I employed a 2 × 2 factorial design to vary whether respondents were told that the proposed map would (A) likely lead to the respondent’s party losing seats in the state legislature, and (B) decrease the likelihood that a co-partisan would represent their district in the state legislature. The first randomly assigned treatment was designed to trigger partisan power considerations, while the second randomly assigned treatment was designed to trigger individual representation considerations. Thus, respondents were assigned to one of four conditions: 1) control (no language after the map introduction); 2) partisan power consideration; 3) legislative representation consideration; 4) both partisan power and legislative representation considerations.Footnote 3 Table 1 displays the experimental vignette language for each treatment condition. Additionally, I included a table at the end of the written treatment condition in the experimental vignette to further clarify the treatments.Footnote 4
Table 1. Experiment #1 treatment conditions
After reading the treatment vignette, respondents evaluated the proposed state legislative district map. First, respondents chose between keeping their state’s current map (status quo) or adopting the proposed map that they read about (positive electoral reform). This question yields a binary measure of respondent support for the proposed electoral reform. Second, respondents rated their support for the proposed map on a seven-point Likert scale, yielding an alternative measure of support for electoral reform.
Experiment #1 results
Using respondents’ choice between the status quo and the proposed state legislative district map as the dependent variable, I explore the extent to which legislative representation considerations and partisan power considerations drive opposition to electoral reform. Panel A in Figure 1 plots the proportion of respondents who chose the newly proposed legislative map over the status quo district map across experimental treatment conditions, along with 95% confidence intervals. The experimental vignette explicitly states that the proposed map improves representation for voters in the respondents’ state, thereby offering a positive electoral reform. Thus, choosing the proposed map reflects democratic principles, and choosing the status quo map over the proposed map represents a departure from democratic principles.
Figure 1. Choice of Map by Treatment Condition.
Note: 95% confidence intervals surround point estimates.
Panel A in Figure 1 reveals a large partisan self-interest treatment effect. Respondents in the control group – those assigned to read an explanation of the proposed map without any additional information on partisan power or individual representation considerations – were slightly more likely to choose the proposed map than the status quo map (51%). However, respondents assigned to any of the three alternative treatment conditions – triggering a partisan power consideration, an individual representation consideration, or both considerations – were far more likely to choose the status quo district map than the proposed map. The percentage of respondents choosing the proposed map over the status quo map is lower than 25% in all three partisan self-interest treatment conditions. Difference of proportions tests confirm that all three of the partisan self-interest treatment conditions decrease the likelihood of choosing the proposed map, relative to the control condition (p-value
$ < $
0.01). Partisan self-interest clearly drives opposition to the proposed electoral reform.
My first hypothesis predicts that partisan power considerations are a more powerful determinant of opposition to redistricting improvements than individual representation considerations. To evaluate this hypothesis, I compare map choice proportions between the middle conditions in panel A – the partisan power consideration and the individual representation consideration. This comparison reveals little difference between the two treatment conditions. Twenty-two percent of respondents who read about the proposed map decreasing their party’s power but leaving their individual representative unchanged chose the proposed map, whereas 21% of respondents who read about the proposed map decreasing their likelihood of being represented by a copartisan but leaving their party’s overall power unchanged chose the proposed map. This difference is statistically indistinguishable from zero and substantially insignificant.Footnote 5 As such, the difference in proportions test yields no evidence in support of hypothesis one.
I also regress the map choice variable on variables for the two treatment conditions using OLS. In Panel B of Figure 1, I display the coefficient estimates generated by this regression, along with 95% confidence intervals. Again, I find that the two treatment conditions have a similar effect on respondents’ choice of legislative district map. The partisan power and individual representation considerations both decrease respondent likelihood of choosing the proposed district map relative to the control condition and to a similar degree.Footnote 6 In other words, I find that respondents are similarly sensitive to electoral loss in the form of individual representation and broad partisan power when it comes to choosing between the status quo and electoral reform.
I further evaluate my first hypothesis by exploring whether Republican and Democratic respondents differ in their responses to the experimental treatments. Partisanship is a central component of public opinion, and recent research suggests Democrats and Republicans differ in their level of support for electoral reform (Coll, Tolbert and Ritter Reference Coll, Tolbert and Ritter2022). This analysis was preregistered as exploratory analysis, meaning I do not offer a hypothesis on whether party affiliation influences treatment effects. In Figure 2 I display the percentage of respondents who chose the newly proposed legislative map across experimental treatment conditions and respondent party affiliation. I classify independent leaners as partisans for this analysis, as leaners are similar to partisans in their support for electoral reform (Coll, Tolbert and Ritter Reference Coll, Tolbert and Ritter2022).
Figure 2. Choice of Map by Treatment Condition and Party Affiliation.
Note: 95% confidence intervals surround point estimates.
The plots in Figure 2 reveal slight differences between partisans. Republican respondents assigned to the control condition and partisan power condition seemingly chose the proposed map at a lower rate than Democratic respondents in the same respective treatment groups. However, the experimental treatments generally had similar effects on Democrats and Republicans in the sample. Regardless of party affiliation, both the partisan power consideration and individual representation consideration led respondents to choose the proposed map at a much lower rate than the control condition.
In sum, I find that the public is sensitive to electoral loss in the form of individual representation and broad partisan power when it comes to choosing between the status quo and electoral reform. The proposed redistricting map in this experiment is explicitly framed as a positive reform from a democratic principles perspective. As such, the results indicate a general propensity to sacrifice principles in favor of partisan self-interest when election results are at stake. I hypothesized above that the partisan self-interest motivation stemmed more from partisan power considerations than individual representation considerations, and I find little support for this proposition. I find a strong partisan power effect in the expected direction, but the individual representation treatment condition led to a surprisingly similar backlash to the proposed district map. Analysis of the alternative dependent variable – support for the proposed map – leads to the same general conclusion, though I find evidence marginally in favor of hypothesis 1. The partisan power consideration leads to slightly less support for the proposed map than the individual representation consideration.Footnote 7 Therefore, I conclude that partisan self-interest can be similarly triggered by broad partisan power considerations and individual representation considerations.
Experiment #2: Loss aversion and principle violation
My second research question further probes the mechanism behind the public’s willingness to violate democratic principles in the name of partisanship: Is partisan gain or partisan loss a more powerful driver of public attitudes on electoral reform? Tapping into the literature on loss aversion and negative partisanship, I argue that partisan loss has more sway over attitudes than partisan gain. I leverage a second survey experiment to test this hypothesis.
Experimental design
I employed a similar research design as the experiment above – an experimental vignette on state legislative redistricting.Footnote 8 I recruited roughly 1,000 respondents from the Lucid Theorem platform, and respondents were again assigned to read about a newly proposed state legislative map.Footnote 9
Respondents had an equal chance of being assigned to each of the following three treatment conditions: 1) neutral – a nonpartisan map that better represents voters in the respondents’ state with no reference to electoral consequences; 2) partisan loss – a nonpartisan map that better represents voters in the respondents’ state but will likely hurt the respondents’ party in upcoming elections; 3) partisan gain – a partisan map that was criticized as unfair but will likely help the respondents’ party in upcoming elections.Footnote 10 Similar to the partisan power consideration treatment in the above experiment, the partisan loss condition makes it clear that the proposed map will hurt the respondent’s party while offering an improvement from a democratic principles standpoint. Conversely, the partisan gain condition is designed to capture the concept of partisan electoral gain at the expense of democratic principles. The experimental language makes it clear that the proposed map will help the respondent’s party while violating democratic principles. Table 2 displays the experimental vignette language for each treatment condition.
Table 2. Experiment #2 treatment conditions
After reading the treatment vignette, respondents chose between keeping their state’s current state legislative district map or adopting the proposed map that they read about. This experiment differs from the first experiment in that the nature of violating democratic principles in favor of partisan self-gain changes based on the treatment condition. One treatment condition (partisan gain) proposed a partisan map that violates democratic principles, while the other two treatment conditions (neutral and partisan loss) proposed a nonpartisan map that follows democratic principles. As such, I developed an additional measure to operationalize the concept of democratic principle violation. I coded respondents in the partisan gain treatment group who chose the partisan proposed map (status quo map) as 1 (0). Conversely, I coded respondents in the neutral and partisan loss treatment groups who chose the status quo map (nonpartisan proposed map) as 1 (0). This measure is better suited to explore the differential impact of partisan gain and partisan loss on respondents’ likelihood of violating democratic principles.
This experimental design builds directly on the first experiment and adds a new dimension to the research by focusing on how citizens respond to the proposition of their party pushing for an unfair electoral reform. The partisan loss reform – giving up partisan power in favor of democratic norms – reflects the decision of majority parties in some states to institute independent redistricting commissions, whereas the partisan gain reform – violating democratic norms in favor of partisan power – reflects the decision of majority parties in other states to engage in partisan gerrymandering. This experiment is designed to clarify whether the opportunity for unfair partisan gain similarly triggers partisan self-interest as the prevention of partisan loss. If loss aversion shapes opinion formation on electoral reform, preventing partisan loss should lead to greater willingness to violate principles than the opportunity for partisan gain.
However, the decision to build directly on the first experiment with the neutral and partisan loss treatment language while incorporating a partisan gain treatment necessitated bundling treatment language. Meaning, there are multiple moving components between the partisan loss and partisan gain treatment conditions. Specifically, the nature of the decision changes from protecting the status quo policy to implementing a new policy. As a result of this treatment bundling, the specific mechanism behind any treatment effects cannot be precisely identified. In other words, while this experiment is capable of determining whether the opportunity for unfair partisan gain and the prevention of partisan loss lead to different levels of willingness to violate democratic norms, it is not capable of specifying the mechanism behind any potential differences.
Experiment #2 results
In Panel A of Figure 3, I display the proportion of respondents who chose the proposed district map over the status quo district map for each of the three treatment groups. Similar to the findings presented above, respondents in the partisan loss group were meaningfully less likely to choose the proposed map than respondents in the neutral condition. Fifty-one percent of respondents in the neutral group chose the proposed map, while 39% of respondents in the partisan loss group chose the proposed map (p-value
$ < $
0.01). Thus, I again find that the threat of partisan loss tends to turn the public against positive electoral reform.
Figure 3. Map Choice by Treatment Condition.
Note: 95% confidence intervals surround point estimates.
While the partisan loss effect is consistent in direction across experiments, its magnitude is substantially smaller in this experiment – 39% support for the proposed map compared to 22% in the first experiment. This result is surprising because the experimental condition language is identical, yet the 95% confidence intervals for the two estimates do not overlap. Differential respondent attentiveness between the two surveys is one potential contributor to the difference in effect size. The first survey applied a slightly more aggressive approach to removing inattentive respondents at the beginning of the survey than the second survey, using two attention screeners instead of one. As a result, the treatment effects in the second experiment (both partisan loss and partisan gain) may be understated; inattentive respondents are equally likely to choose the proposed map and status quo map, yielding conservative estimates of experimental effects.Footnote 11
Surprisingly, only 34% of respondents in the partisan gain condition, who were introduced to a map that would unfairly help their own party, chose the proposed map. Relative to the status quo, respondents were more likely to choose an outcome-neutral map that improves representation for one’s state (neutral group) than choose a map that unfairly helps one’s own party (p-value
$ < $
0.01). In sum, panel A shows that respondents were not eager to support their party’s attempt to gain an unfair electoral advantage.
As mentioned above, the proposed map in the partisan gain condition differs in substance from the proposed map in the partisan loss and neutral conditions. The partisan gain condition proposes a partisan map that violates democratic principles, while the other two conditions propose a nonpartisan map that follows democratic principles. As such, I developed a new variable to indicate the violation of democratic principles that incorporates this difference. Respondents in the partisan gain treatment group who chose the proposed map over the status quo are coded as 1 because the proposed map is described as not accurately reflecting voters in the respondent’s state. Conversely, respondents in the neutral and partisan loss treatment groups who choose the status quo over the proposed map are coded as 1, because the proposed map is described as more accurately reflecting voters in the respondent’s state. In Panel B of Figure 3, I plot the proportion of respondents in each treatment group who violated the democratic principle cue in favor of partisan self-interest. In other words, this plot draws on the same data as panel A but provides a more relevant classification for the research question under inquiry.
In Panel B of Figure 3, I show that the partisan gain and partisan loss treatments have different effects on respondents willingness to violate democratic principles in favor of partisan self-interest. More than 60% of respondents assigned to the partisan loss condition chose partisan self-interest over explicitly stated democratic norms by selecting the status quo map. Conversely, less than 40% of respondents assigned to the partisan gain condition chose partisan self-interest over explicitly stated democratic norms by selecting the partisan map. Relative to the neutral condition, the prospect of a fairer map that leads to partisan loss made respondents more likely to violate democratic principles (p-value
$ < $
0.01), and the prospect of an unfair map that leads to partisan gain made respondents less likely to violate partisan self-interest (p-value
$ < $
0.01). In other words, respondents tended to reject fair redistricting reform that would lead to partisan loss, but they also rejected unfair redistricting reform that would lead to partisan gain. Therefore, I report evidence in support of hypothesis 2 – citizens are more willing to sacrifice principles to prevent their party from losing power than to unfairly help their party gain power.Footnote
12
I further explore the partisan gain and partisan loss treatment effects by sub-setting the analysis by respondent party affiliation. In Figure 4, I display the percentage of respondents who violated democratic principles across experimental treatment conditions and respondent party affiliation. I again classify independent learners as partisans. Figure 4 shows that, directionally, democratic and republican respondents responded similarly to the treatment conditions. For both partisan subsets, the prospect of fair partisan loss led to greater principle violation than the prospect of unfair partisan gain.
Figure 4. Democratic Principle Violation by Treatment Condition and Partisanship.
Note: 95% confidence intervals surround point estimates.
The results from this experiment follow the expectations of Hypothesis 2, but it is important to note that there are multiple moving components between the partisan loss and partisan gain treatment conditions. In addition to the difference between partisan loss and partisan gain, the nature of the decision changes from protecting the status quo policy to implementing a new policy. As a result of this treatment bundling, the specific mechanism behind the partisan gain treatment effect remains unclear. I use the concept of loss aversion to argue that the public is less willing to violate principles to attain partisan gain than to prevent partisan loss. Alternatively, the mechanism behind the observed effect might stem from a preference for the status quo over the new policy. Status quo bias is both a product of loss aversion and an independent psychological force (Eidelman and Crandall Reference Eidelman and Crandall2009; Samuelson and Zeckhauser Reference Samuelson and Zeckhauser1988; Tversky and Kahneman Reference Tversky and Kahneman1991). A status quo bias, independent of loss aversion, might serve as the primary mechanism behind the experiment effect.
In fact, respondents consistently displayed a desire to retain the status quo over a new district map in both experiments. The control (neutral) condition, which presented a positive cue on democratic norms with no electoral downside for the respondents’ political party, only yielded around 50% support for the proposed map in both experiments. Further, all treatment conditions resulted in even less support for the proposed map, ranging between 20% and 40%. The primary finding from the first experiment – that partisan power and individual representation considerations similarly activate resistance to principled electoral reform – holds significance irrespective of the underlying mechanism. However, the findings across both experiments suggest that status quo bias might be a more powerful driver of opinion on electoral reform than loss aversion. In sum, while the evidence presented here supports the loss aversion hypothesis, further testing is needed to more clearly flesh out why preventing partisan loss is a more powerful driver of Americans’ willingness to violate democratic principles than achieving partisan gain.
Conclusion
At a time of hostile partisanship and concerns over democratic backsliding, clarifying the determinants of public support for electoral reform is critical for understanding the electoral landscape and partisan battleground in the United States. Partisan self-interest plays a key role in the formation of mass opinion on electoral reforms (Alarian and Zonszein Reference Alarian and Zonszein2024; Biggers Reference Biggers2019; Biggers and Bowler Reference Biggers and Bowler2022; Bowler and Donovan Reference Bowler and Donovan2016; Kane Reference Kane2017), but principles and core values also influence this process (Biggers and Bowler Reference Biggers and Bowler2023; McCarthy Reference McCarthy2022; Virgin Reference Virgin2023). I advance this stream of research by investigating the conditions under which the public is more or less likely to prioritize partisan self-interest over principles.
Leveraging two survey experiments, this research offers two meaningful contributions to the study of when and how the public is willing to violate principles in favor of partisan self-interest. First, I show that both partisan power and individual representation are powerful enough considerations to activate partisan self-interest in the formation of public opinion on electoral reform. Second, I demonstrate an asymmetry between preventing partisan loss and achieving partisan gain as triggers of democratic principle violation. The threat of partisan loss is a powerful incentive for the violation of principles concerning electoral reform, whereas the opportunity for partisan gain is not.
This research is limited in important ways, leading to opportunities for additional inquiry. First, questions remain on the generalizability of the findings presented above. Presenting respondents with actual reform proposals rather than hypothetical proposals would improve the external validity of this research. Additionally, the experimental designs rely on respondents lacking precise knowledge of redistricting processes and partisan power structures in their state. If a respondent knows that their state legislature or state legislative district is extremely noncompetitive, the threat of electoral loss is muted. While studies suggest a generally low level of citizen knowledge of state politics (Carpini and Keeter Reference Carpini and Keeter.1996; Lyons, Jaeger and Wolak Reference Lyons, Jaeger and Wolak2013; Panagopoulos Reference Panagopoulos2013), future research should focus specifically on states where electoral reforms are plausible and impactful on partisan power structures. Further, this study does not take respondent knowledge of redistricting into account, yet such knowledge plausibly conditions how respondents evaluate reform proposals. Recent research highlights the critical role of redistricting knowledge in shaping public opinion on redistricting processes (McLaughlin et al. Reference McLaughlin, Olson, Barron, Benjamin Ashton, Blum, Finocchiaro and Crespin2024; Panagopoulos Reference Panagopoulos2013). Further research should evaluate the extent to which redistricting knowledge conditions the experimental effects displayed above.
Second, this study focuses on a single type of reform, a proposed legislative district map, with a relatively minor normative signal on democratic principles. Future research can expand on the findings presented above by exploring how voters consider different redistricting methods. For instance, the normative signal can be strengthened to give a more explicit choice between following and breaking democratic norms.
Third, further research is needed to flesh out the causal mechanism behind the loss aversion finding reported above. Results suggest the public deferentially weighs prospective partisan gains and losses in opinion formation on electoral reforms. However, the multiple moving components of the experimental treatments mean that alternative explanations for the observed treatment effect, such as status quo bias, cannot be ruled out. Further research should develop new ways to test for loss aversion in public opinion on electoral reform that allows for stronger statements on the causal mechanism.
Finally, I limit the reform under question to redistricting. While the question of partisan power versus individual representation is particularly relevant to redistricting reforms, the concept of loss aversion plausibly applies to electoral reform more broadly. Further research is needed to determine the power of loss aversion in shaping public opinion on elections and electoral reform.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/spq.2024.34.
Funding Statement
This research was funded by the American Political Science Association (Reference Number: AGR10172022). The financial sponsor played no role in the design, execution, analysis and interpretation of data, or writing of the study.
Acknowledgements
I thank Adam Joseph Ramey and Joseph Coll for their helpful suggestions.
Financial support
This research was made possible by funding from the APSA Doc toral Dissertation Research Improvement Grant (DDRIG) Program.
Conflict of Interest
The author declares no potential conflicts of interest with respect to the research, authorship, and/or publicationof this article.
Author biography
Peter McLaughlin is an Assistant Professor of Political Science at the University of Rhode Island. His research agenda focuses on congressional representation, distributive politics, and elections.
Open access
