Information and Issue Constraints on Party Cues

Abstract

Past research has found that citizens will support either side of a policy debate if their party endorses it, regardless of the policy details. Such results cast doubt on the electorate’s ability to direct and constrain public officials. Yet other studies find that people give priority to policy information in their decision-making. We hypothesize that the relative effects of party cues and policy details depend on the degree to which people can identify the ideological direction of a policy. Using a survey experiment, we show that for policies readily classified as liberal or conservative, preferences are less influenced by party cues than by policy details. Public policy information has the greatest impact on preferences relative to party cues when the information establishes the ideological direction of the policy by indicating the values at stake and the groups that will be helped or hurt by the policy. Information therefore is most impactful on policies that are ideologically ambiguous in the absence of policy details.

Keywords

information party cues motivated reasoning heuristics policy preferences

For decades, scholars have recognized that partisanship exerts considerable influence on public opinion and voting in the American electorate. Campbell, Converse, Miller, and Stokes (1960) assert that, “identification with a party raises a perceptual screen through which the individual tends to see what is favorable to his partisan orientation” (p. 133). Following party cues can be an efficient heuristic (Popkin, 1991), but it can also be costly if it causes people to ignore information and support policies and candidates they would oppose in the absence of partisanship. When partisans engage in motivated reasoning, they seek-out information to bolster their party’s policy positions and argue against information that challenges their party. As a result, citizens are less affected by the merits of policy arguments than by their partisan sources (e.g., Bolsen, Druckman, & Cook, 2014).

Experimental studies have shown that party elites can adopt widely varying positions on political issues without risking public support. Cohen (2003) found that people are persuaded to favor radically different social welfare policies if they originate from their own party. Similarly, Rahn (1993) and Riggle, Ottati, Wyer, Kuklinski, and Schwarz (1992) discovered that respondents continued to support their party’s candidates even when they strayed from the party line on core issues. Further evidence of the strength of party cues has been documented in other policy realms (Druckman, 2001; Kam, 2005), public acceptance of Supreme Court decisions (Nicholson & Hansford, 2014), and decision-making outside of politics (Iyengar & Westwood, 2015).

Yet other research identifies instances in which people pay attention to substantive information in making judgments. For example, individuals blamed officials of the opposing party for problems in the wake of Hurricane Katrina, but made less partisan attributions when informed about the officials’ job responsibilities (Malhotra & Kuo, 2008). Party cues have been found to have little effect on attitudes toward a salient issue like abortion (Arceneaux, 2008), and policy information appears to have larger effects than party cues on preferences when people are given extensive descriptions of an issue (Bullock, 2011). Furthermore, one-sided presentations of policy information have been shown to influence people’s preferences on ballot initiatives beyond the effects of party (Boudreau & MacKenzie, 2014).

These conflicting findings raise the question of how much latitude parties enjoy in the positions they take before their followers will abandon them. In this article, we examine how variation in the issues and presentation of policy details affects the relative impact of party cues and policy information on preferences. First, we believe the features of issues analyzed are consequential, in particular the extent to which a policy has a recognizable ideological structure. Some policies are more readily classified by the public as liberal or conservative proposals because of the values they further and the groups that benefit from them. On such issues, we expect people will rely less on party cues and give more weight to the substance of the policy, whereas party cues should be more influential on policies that are ideologically ambiguous. Therefore, we assess how the effects of cues and information differ across multiple issues deliberately selected to provide variation in the clarity of their underlying ideological divisions.

Second, the addition of policy information should have the greatest impact on issues that are initially difficult to interpret. Policy information that sharpens the ideological divisions on an issue can increase the likelihood that opinions will be guided by the substantive features of the policy beyond the influence of partisan cues. Therefore, we test how adding policy information through elite debate about the consequences and beneficiaries of a policy can affect preferences and moderate the influence of party cues.

We present results from a survey experiment in which people evaluate several public policy proposals following exposure to different combinations of party cues and policy information. Our results indicate that although party endorsements influence policy preferences, such influence is muted on policies that have a more evident ideological structure. Exposure to information about a policy’s consequences and beneficiaries conveyed through policy arguments can have independent effects that rival and exceed the impact of party cues. Notably, the effects of party cues and arguments are not uniform, but rather are contingent on the issue. These results indicate there are identifiable circumstances in which information penetrates the partisan “perceptual screen” and that citizens can weigh substantive policy details alongside their party loyalties.

The Features of Issues and Policy Information

While it seems intuitive that the impact of partisan cues may vary across issues, we lack a theoretical understanding and direct tests of the specific features of issues that can override partisanship. People presumably have stronger attitudes on salient issues and require less guidance from party leaders on them (e.g., Ciuk & Yost, 2016). However, several studies have also found that party cues can be dominant on salient issues, including core domestic and foreign policy issues that reliably separate Democrats and Republicans and affect a variety of familiar group beneficiaries (Cohen, 2003; Rahn, 1993; Riggle et al., 1992).

Differentiating among these studies requires additional testing of the features of issues that condition the impact of policy details and partisan cues. We believe a critical dimension is whether the issue has a clear ideological structure. Knowing the interests and values that are furthered by a policy increases the likelihood that people can classify the policy as liberal or conservative. Policies that, for example, address economic inequality, increase government spending on social programs, provide assistance to the working class and unemployed, or benefit racial and ethnic minorities indicate liberal policy goals while policies that favor the affluent, give priority to business interests, or reduce taxes and government services and programs signal conservative goals.¹

Some issues, such as abortion rights, have been in the public consciousness for such a long time that most people know that policies restricting abortion rights are conservative policies. On such issues, we expect that people will depend less on party cues to decide where they stand on the policy. Yet not all issues are so transparent that people automatically recognize which side of the issue is the liberal or conservative position. In our view, additional information is most effective relative to party cues when it brings out the ideological direction of a policy by identifying the values at stake and the segments of society that will be helped or hurt by it.

More information by itself (cf. Bullock, 2011) therefore will not always increase the weight of policy details if the information does not clarify the ideological direction of the issue. If we are correct in this assumption, an economical description of a policy that maps onto people’s ideological values can be highly effective. Similarly, salience may be neither sufficient nor necessary to increase the effect of policy details (cf. Ciuk & Yost, 2016). Salient policies may be more likely to have well defined polarities that correspond to people’s ideological values, but less salient policies can also possess or acquire these features.

Certainly, a complete explanation of the relative impact of party cues and policy information needs also to encompass the strength of party cues and the motivations and capacities of individuals,² but here we focus on what we believe is the most notable omission in existing work: specifically (a) the effect of issues that vary in the clarity of their ideological divisions, and (b) the effect of political arguments that emphasize the implications and group beneficiaries of a policy.

Hypotheses

Our hypotheses are guided by the assumption that every cue and piece of information (e.g., the substance of the issue, party positions, and arguments for and against the policy) potentially affects how people construe the purpose and desirability of a policy. The marginal effect of adding an endorsement or policy details depends on how informed citizens already are about the policy. As a general proposition, given low average information levels in the electorate (Delli Carpini & Keeter, 1996), we expect that people will rely on party cues for guidance on what positions to adopt on public policies. Therefore, our first hypothesis is:

Hypothesis 1a (H1a): Party endorsements of a policy will increase policy support among party identifiers.

Yet we think the nature of the issues analyzed will condition the role of party cues. People will depend more on their party when they are unable to discern and evaluate the implications of the policy. This leads to a refinement of H1a:

Hypothesis 1b (H1b): Party cues will have a larger effect on issues with less obvious ideological divisions than on issues with a clear ideological structure.

The interaction between party cues and issues should be particularly evident when party positions are “reversed” or contrary to stereotypes, such as when Democrats endorse a conservative policy and Republicans favor a liberal policy (e.g., see Cohen, 2003; Rahn, 1993). The public will be more likely to follow reversed partisan cues when it has difficulty comprehending the ideological direction of the policy.

We test two hypotheses about the effects of policy information. H2 tests whether adding ideological structure to the issues, by providing substantive arguments about the group interests and values affected by the policy, changes preferences:

Hypothesis 2 (H2): Arguments that provide ideological structure to an issue will reinforce traditional party cues, but offset reversed party cues. As with party cues, political arguments will have a greater impact on unstructured issues than on issues that are already defined ideologically.

H3 examines the effect of policy details by testing how sensitive preferences are to the ideological direction of the policy, all else constant. We hypothesize the policy effect (i.e., the effect of changing the policy direction from liberal to conservative) is conditional upon the degree to which respondents are aware of its ideological alignment:

Hypothesis 3 (H3): The policy effect will be greatest on issues that pose clear ideological alternatives, and weakest on issues that are not well defined ideologically.

It follows that on policies that respondents have difficulty classifying ideologically, exposure to arguments about the implications of the policy will sharpen perceptions of the contrasts between competing positions. Therefore, our final hypothesis is:

Hypothesis 4 (H4): Switching the policy direction will produce a larger policy effect when people are informed about policy consequences than when they are not.

Research Design

To test these hypotheses, we use a survey experiment that varies three features of political information. First, we vary the presence or absence of party endorsements, and if present, whether the party endorsement is the party’s traditional policy position (i.e., its position in reality) or the reverse of its real position. Consistent with past studies (Cohen, 2003; Mullinix, 2016, 2018), we use this design to assess the degree to which people follow their party’s lead despite contradictory policy information. Second, we vary whether people receive arguments containing information that clarifies the effects and beneficiaries of a policy. Such clarifying arguments potentially reinforce party positions, but they can also conflict with party cues. Finally, we varied the ideological direction of the policy proposal relative to the status quo. For each policy issue, we constructed a liberal and a conservative version to test whether people are more responsive to the ideological direction of a policy or to the party endorsing it. The 3 × 2 × 2 design yields the 12 experimental conditions shown in Table 1.

Table 1.

Experimental Conditions.

	Liberal proposal		Conservative proposal
	No arguments	Arguments	No arguments	Arguments
No party cues	Group 1	Group 2	Group 3	Group 4
Traditional party cues	Group 5	Group 6	Group 7	Group 8
Reversed party cues	Group 9	Group 10	Group 11	Group 12

Issue Selection

We built treatments combining party cues and policy arguments around liberal and conservative versions of issues pertaining to taxes, preschool funding, and unemployment benefits. We selected these three issues because they were not equally recognizable in pretests as liberal or conservative policies; that is, they differed in the clarity of their ideological divisions. Such variation across issues allows us to test the relative impact of party cues and policy information when ideological divisions on an issue are either easily discerned or obscure.

The tax issue is a relatively unfamiliar issue concerning whether states should place greater priority on state sales taxes or state income taxes in generating revenue. The conservative version of this policy calls for greater reliance on state sales taxes, while the liberal version calls for increased reliance on state income taxes. The preschool funding issue concerns whether the federal government should provide aid (liberal version) or rescind aid (conservative version) to states for prekindergarten classes. Finally, the unemployment benefits issue involves conventional proposals to either reduce (conservative version) or increase (liberal version) the amount and duration of unemployment benefits.

To assess perceptions of the ideological structure of these issues, we implemented pretests in which respondents evaluated the ideological direction of each policy.³ Results from these pretests reveal substantial differences in the ideological classifications of the three issues. The unemployment issue had the clearest ideological structure for respondents. When no party cues were provided, 72% of respondents identified cuts to unemployment benefits as a conservative policy and only 13% labeled it as a liberal policy. There was less consensus on the preschool issue, as 58% believed that increasing federal funding for preschool programs was a liberal policy while only 13% viewed it as conservative. Finally, as we expected, the tax issue generated the greatest ambiguity among respondents, with 37% believing a shift toward sales taxes is a conservative policy, and 30% judging it to be a liberal proposal. The pretests therefore confirm these three policies do not possess equally discernible ideological divisions and, therefore, the effect of party cues and substantive information should vary in theoretically predictable ways across these issues.⁴ The issue with the least ideological clarity—sales taxes versus income taxes—should be most susceptible to party and information effects, while the issue with the greatest ideological clarity—unemployment benefits—should be relatively impervious to party cues and arguments.

Sample and Stimuli

The experiments were implemented with an online convenience sample using Amazon Mechanical Turk (MTurk). While not a probability-based sample of the national population,⁵ experiments implemented simultaneously in MTurk and with nationally representative samples often produce indistinguishable treatment effects (Coppock, 2018; Mullinix, Freese, & Druckman, 2015). Clifford, Jewell, and Waggoner (2015) also demonstrate that the values and psychological characteristics of liberals and conservatives in MTurk mirror those in the mass public. While more diverse than other common experimental samples (e.g., students), a limitation of MTurk is that it is disproportionately Democratic (our sample contains 899 Democrats and 320 Republicans), yielding less statistical power for analyses of Republicans (see Supplementary Materials).

The experimental stimuli were presented in the form of news articles (one article for each issue). Respondents were randomly assigned to one of the 12 conditions (randomization checks are shown in Supplementary Materials). The order of the issues was fixed. Respondents first received an article on the tax proposal, followed by an article on preschool funding, and then by an article on unemployment benefits. No spillover effects stemming from the order of the issues were detected by our tests.⁶ After respondents read each article, they were asked a series of questions evaluating their policy preferences and perceptions of the issue.

The stimuli were based on the wording of actual news articles to maximize external validity (see Supplementary Materials for the treatment text). For each issue, we created baseline control conditions for the liberal and conservative policy that provided a simple description of the policy but offered no party cues or rationale for or against the policy. This condition establishes how respondents evaluate the features of a policy without any additional guidance. For example, the basic description of the conservative version of the tax policy is:

Despite opposition, several governors across the country have proposed to increase reliance on state sales taxes while cutting income taxes, setting up ambitious experiments in tax reform. The idea is to lower state income taxes and make up the lost revenue by increasing state sales taxes.

The liberal version is identical except it reverses the priorities by calling for an increase in income taxes:

Despite opposition, several governors across the country have proposed to increase reliance on state income taxes while cutting sales taxes, setting up ambitious experiments in tax reform. The idea is to lower state sales taxes and make up the lost revenue by increasing state income taxes.

The format of the other two issues is similar. The baseline for the preschool policy described proposals either to fund (liberal) or to end funding (conservative) for a US$30 billion program giving federal aid to states. The baseline for the unemployment issue described proposals either to increase (liberal) or cut benefits (conservative) for the unemployed.

In the traditional party endorsement conditions, the Democratic Party endorsed the liberal positions on the policies while the Republican Party was reported as taking the conservative positions. These party endorsements were simply reversed on each policy in the reversed party cue conditions.

In addition to manipulating party cues, we varied the presence or absence of information that identifies the beneficiaries and ideological values of each policy. The policy information is conveyed through arguments of roughly equal length for and against each proposal. For example, on the tax proposal, supporters of the conservative version state the policy “will simplify the tax system, make their states more competitive in attracting employers and high-skilled workers, and curb pressure for more government spending” but opponents argue that the policy will “increase economic inequality by reducing taxes predominantly on the affluent” and lead to “cutbacks in education and vital social services.”

On the early childhood education issue, proponents assert that legislation will help “make preschool available for low-income children” and promote “equal opportunity.” Opponents contend the program reflects the “inefficiencies of big government” and “invites too much federal government control over how states and local communities run preschool programs.” Supporters of reduced unemployment benefits claim the cuts are needed to “help pay down the state’s debt and improve business conditions for economic growth.” Opponents assert that a reduction in benefits “cuts a large hole in the government safety net” and would hurt “African American and Hispanic workers [who] have been hit hardest by the downturn.”⁷

Measures

The primary dependent variable of interest—policy support for each proposal—was asked immediately after the participant read each article, with responses ranging from 1 (strongly oppose) to 7 (strongly support).

To determine whether there was evidence of disconfirmation bias and counter arguing, participants who were assigned to treatments that included substantive arguments rated the effectiveness of those arguments (see Taber, Cann, & Kucsova, 2009). To understand how the treatments shaped people’s understanding of the ideological alignment of the proposals, we asked people to evaluate the extent to which each proposal was either liberal or conservative.

Results

Our pretests indicated respondents had the greatest difficulty identifying the ideology of competing sides of the tax policy, while changing unemployment benefits was viewed as the most ideologically transparent. These pretest results are largely confirmed in the experiment although there are important differences between Democratic and Republican respondents that allow us to refine our hypotheses. The results in Appendix Tables A1 and A2 show whether respondents perceived ideological differences between competing proposals on each of the three issues. These regressions are based only on respondents who were assigned to conditions in which they received neither party cues nor arguments. Thus, when the policies are described without embellishment, Democrats correctly perceived the conservative tax policy to be significantly more conservative than the liberal tax proposal but the difference is just under a point on the 7-point scale (β = 0.937). In contrast, the differences in their ideological classifications of competing versions of the preschool (β = 2.483) and unemployment policies (β = 3.654) are more than twice as large, confirming they have no difficulty distinguishing liberal from conservative proposals in these two policy realms.

Republicans also see the largest ideological difference between competing versions of the unemployment policy (β = 2.662). But they perceive a slightly larger contrast between the liberal and conservative versions of the tax policy than do Democrats (β = 1.387), and observe the smallest contrast on the preschool issue (β = 1.247). Moreover, on the tax and preschool issues, Republicans appear to have difficulty classifying the conservative versions of each policy. On both issues, they perceive the liberal policy to be solidly liberal, with regression constants around 3.0 on a 7-point scale ranging from very liberal (1) to very conservative (7). However, they perceive neither the proposal to give priority to sales taxes nor the policy to defund the preschool program as being especially conservative in the baseline condition, rating it on average only slightly above the 4.0 midpoint.

These results on the ideological clarity of the three issues suggest the following implications:

The tax issue should have the greatest latitude, especially among Democrats, to be redefined by party endorsements and information about the implications of the policy.

By the same token, the ambivalence of Republicans toward the ideology of the preschool policy in the baseline condition should make Republican respondents more susceptible to persuasion by party cues and information on that issue.

In contrast, manipulating party cues or elaborating on the consequences of the unemployment policy should have little impact on either Democrats’ or Republicans’ preferences, because respondents already know enough from the baseline description of the issue to have a clear preference toward the policy.

Our analysis proceeds as follows: Policy support (the dependent variable) is regressed on the experimental conditions for each of the three issues. For each regression model, the baseline condition is the no party cue, no argument condition. The regression coefficients therefore represent the change in policy support—on the 7-point scale—produced by each experimental manipulation, relative to the no party cue, no argument baseline. Positive coefficients indicate an increase in policy support, and negative coefficients reveal a decrease in support. Separate models are reported for the liberal and conservative versions of each policy, and because the relationship between the respondent’s partisanship and elite party cues is central to our predictions, we also break all results down by the partisanship of respondents (see Bolsen et al., 2014).⁸ Throughout our discussion, we will report differences between selected conditions (i.e., average treatment effects); all estimates and treatment effects are captured by the regression coefficients.⁹

Democratic Respondents

Tax policy

Although we generally expect respondents to be guided by party cues (H1a), we hypothesize such cues will have a larger influence on Democratic respondents’ tax policy preferences when no additional information is provided because this issue was the most opaque of the three issues for them (H1b). Table 2 largely supports this hypothesis as respondents’ preferences consistently move in the direction of the party cue, although only one of the four coefficients is statistically significant. Democrats are slightly more supportive of increasing income taxes when their party backs this liberal policy (β = 0.261), but they become significantly less supportive when told the Democratic Party opposes this policy (β = −0.836). Likewise, they become slightly less supportive of the conservative tax proposal when told their party opposes it (β = −0.05), and increase their support when informed that Democrats propose it (β = 0.384).

Table 2.

Policy Support (Democrats Only).

	Liberal tax	Conservative tax	Liberal preschool	Conservative preschool	Liberal unemployment	Conservative unemployment
Traditional party cues	0.261	−0.0484	0.381	0.0690	−0.129	0.0515
Traditional party cues	(0.249)	(0.264)	(0.259)	(0.294)	(0.277)	(0.257)
Reversed party cues	−0.836***	0.384	0.184	0.537*	−0.00809	0.493*
Reversed party cues	(0.248)	(0.269)	(0.255)	(0.302)	(0.275)	(0.255)
Arguments	0.647**	−0.721***	0.579**	0.567*	−0.0337	0.370
Arguments	(0.252)	(0.259)	(0.260)	(0.304)	(0.276)	(0.253)
Traditional cues + Arguments	0.826***	−1.043***	0.304	−0.00463	−0.321	−0.0940
Traditional cues + Arguments	(0.246)	(0.266)	(0.256)	(0.302)	(0.281)	(0.263)
Reversed cues + Arguments	0.139	−0.282	0.722***	0.829***	−0.0294	0.220
Reversed cues + Arguments	(0.245)	(0.257)	(0.255)	(0.302)	(0.269)	(0.257)
Constant	4.086***	3.733***	4.893***	2.519***	5.115***	2.494***
Constant	(0.179)	(0.185)	(0.182)	(0.207)	(0.204)	(0.178)
Observations	451	448	452	447	432	467
R ²	.115	.078	.023	.030	.005	.016

Note. Standard errors in parentheses.

p < .1, two-tailed. **p < .05, two-tailed. ***p < .01, two-tailed.

We next test H2 to see whether party cue effects are sustained when respondents are exposed to arguments that inform them of the ramifications of the policy. The models in Table 2 show that arguments containing clarifying policy information have effects that rival and sometimes exceed the effects of party cues. Arguments alone, without party cues, significantly increase support for the liberal proposal (β = 0.647), and significantly reduce support for the conservative proposal (β = −0.721). The combination of traditional party cues and arguments polarizes judgments of the two policies even further (Liberal Tax β = 0.826; Conservative Tax β = −1.043). While Democrats express only a slight preference for the liberal tax proposal over the conservative proposal when aided solely by party cues (Δ = 0.66, p < .007), they prefer the liberal version of the tax proposal by a wide margin (Δ = 2.22, p < .001) when arguments augment party cues.

The power of tax policy information conveyed by arguments is most apparent in the reversed party cues condition, when the arguments and party cues push in opposing directions. Without the substantive information contained in arguments, Democrats are led astray by the reversed party cue to become significantly more opposed to the liberal proposal (β = −0.836), but this effect vanishes when arguments are presented alongside the reversed cue (β = 0.139). Similarly, the reversed party cue prompts Democrats to become more supportive of the conservative proposal (β = 0.384), but adding arguments causes them to become less supportive (β = −0.282). When party cues are reversed without arguments, Democrats actually prefer the conservative tax proposal to the liberal proposal (Δ = 0.87, p < .001), but the addition of arguments corrects this reversal and reestablishes the liberal proposal as the more preferred policy (Δ = 0.77, p < .001). Therefore, with the aid of greater policy information, Democrats clearly prefer the liberal policy (increasing income taxes) to the conservative policy (increasing sales taxes) regardless of the Democratic Party position on the proposals.

The influences of party, arguments, and policy direction are interrelated. When policy information is provided through arguments, respondents see greater contrast between competing versions of the policy, and preferences therefore are more substantially affected by a change of policy direction. Without substantive arguments, Democrats express little difference in preference between liberal and conservative versions of the tax policy when both are either endorsed (Δ = 0.23, p < .194) or opposed (Δ = 0.43, p < .042) by the Democratic Party.¹⁰ In contrast, as predicted by H4, once the impacts of the competing tax policies are detailed in opposing arguments, the ideological direction of the policy has a large effect on preferences when the party cue is held constant. Adding substantive information in the form of arguments clarifies the ideological implications of the policy for Democratic respondents and magnifies the effect of changing the policy direction from liberal to conservative (Δ = 1.46, p < .001 when the Democratic Party supports both policies; Δ = 1.53, p < .001 when the Democratic Party opposes both policies).

Unemployment benefits

The unemployment issue stands in sharp contrast to the tax issue. Table 2 shows Democrats require no assistance to comprehend and evaluate this highly ideological issue. As predicted by H1b, party cues have little effect on preferences when people already understand the implications of the policy. Only the reversed party cue’s effect on evaluations of the conservative proposal is significant (β = 0.493). Consistent with H2, arguments about the consequences of changing the level and duration of unemployment benefits also do not significantly move preferences. But, in accord with H3, the policy direction has a huge effect on preferences even without arguments or party cues because the unemployment issue is highly structured with clear policy implications (the difference between the constant coefficients is 2.62, p < .001).

Preschool funding

The dynamics of the preschool issue follow a pattern similar to the unemployment issue. Democratic respondents strongly prefer the liberal version over the conservative version of the policy in the baseline condition without party cues or arguments (Δ = 2.37, p < .001). Party cues create only modest movement of policy preferences (supporting H1b)—traditional cues slightly, but not significantly, increase support for the liberal policy (β = 0.381), while reversed cues significantly increase support only for the conservative policy (β = 0.537).

The effects of arguments are less consistent and smaller in magnitude on the preschool issue than on the tax issue. Unexpectedly, in the baseline condition, arguments highlighting the preschool policy’s beneficiaries significantly increase support for both the liberal and conservative policy versions (Liberal Preschool β = 0.579, Conservative Preschool β = 0.567). However, arguments do not move preferences significantly when traditional party cues are present. Arguments do have a corrective effect by significantly increasing Democrats’ support for the liberal preschool policy in the reverse party cue condition when the Democratic Party is represented as being against the policy (β = 0.722); but arguments fail to offset the reversed party cue effect on the conservative proposal (β = 0.829).

The implications of the preschool policy are sufficiently transparent that changing the direction of the policy has a large effect on preferences across all conditions (consistent with H3). As noted above, the difference in support for the liberal and conservative policies, at baseline, exceeds 2.0 points in the expected direction on the 7-point scale. Democrats have consistent preferences across conditions on this issue in favor of the liberal policy.

Republican Respondents

Tax policy

Republicans respond somewhat differently to the tax policy compared with Democrats. Like Democratic respondents, Republican respondents, shown in Table 3, shift their preferences on the tax issue according to party endorsements. Although the traditional party cue does not significantly change preferences, the reversed party cue significantly raises Republicans’ evaluation of the liberal policy (β = 0.793) and significantly diminishes their support for the conservative policy (β = 0.802).

Table 3.

Policy Support (Republicans Only).

	Liberal tax	Conservative tax	Liberal preschool	Conservative preschool	Liberal unemployment	Conservative unemployment
Traditional party cues	−0.485	−0.461	−0.966**	0.220	−0.909*	−0.110
Traditional party cues	(0.472)	(0.486)	(0.484)	(0.559)	(0.485)	(0.522)
Reversed party cues	0.793*	−0.802*	0.465	−0.548	−0.554	0.243
Reversed party cues	(0.463)	(0.428)	(0.538)	(0.540)	(0.447)	(0.517)
Arguments	0.150	−0.689	−0.0288	−0.0518	−0.587	0.0373
Arguments	(0.482)	(0.474)	(0.484)	(0.519)	(0.447)	(0.539)
Traditional cues + Arguments	0.122	−0.262	−1.071**	0.0952	−0.354	0.321
Traditional cues + Arguments	(0.467)	(0.435)	(0.519)	(0.559)	(0.463)	(0.512)
Reversed cues + Arguments	0.673	−0.800*	−0.514	0.406	−0.338	0.201
Reversed cues + Arguments	(0.494)	(0.465)	(0.514)	(0.536)	(0.499)	(0.546)
Constant	3.100***	5.080***	4.154***	3.905***	3.909***	4.571***
Constant	(0.322)	(0.328)	(0.360)	(0.408)	(0.311)	(0.362)
Observations	157	163	160	160	166	154
R ²	.057	.033	.081	.026	.025	.007

Note. Standard errors in parentheses.

p < .1, two-tailed. **p < .05, two-tailed. ***p < .01, two-tailed.

Policy arguments on the tax issue fail to completely counteract the influence of reversed party cues among Republicans (as they do among Democratic respondents). This is contrary to H2 and can be explained by examining how Republicans evaluate the arguments on this issue (see the regressions on the evaluation of arguments in the Supplementary Materials). Arguments can keep errant party cues in check only if Republicans find the information contained in the conservative argument sufficiently compelling that they will support the conservative policy despite Republican opposition to it. Instead, what we find in the reversed cue condition is that Republicans judge the conservative argument to be no stronger than the liberal argument. This contrasts with how Democrats consistently rate the liberal argument as being significantly stronger than the conservative argument across all conditions, which causes them to favor the liberal tax policy regardless of party cues. In the parlance of framing studies, the arguments on the tax policy resonated more strongly among Democrats than Republicans (Chong & Druckman, 2007). A different set of arguments therefore might have offset party cues to a greater degree on this issue among Republicans.

Although a Republican endorsement can marginally increase support among Republican respondents for either version of the policy, the ideological direction of the policy is the dominant influence on preferences on the tax issue. Republican respondents prefer by a significant margin the conservative proposal to the liberal proposal regardless of where the parties stand on the issue. Moreover, adding substantive arguments about the consequences of each policy does not consistently magnify the influence of the policy direction.¹¹

Preschool funding

The preschool issue is a less structured issue for Republican respondents than is the tax issue. Republican respondents are indifferent between the liberal and conservative policies in the baseline conditions (Δ = 0.25, p < .34). It is therefore not surprising that without policy details, party cues have more latitude to move preferences in the expected directions, as predicted by H1b. Traditional party cues significantly reduce Republicans’ support for the liberal policy and marginally increase their support for the conservative policy. Reversed party cues increase Republicans’ support for the liberal policy and reduce their support for the conservative policy. Although the effects of the reversed party cues are not statistically significant, they are sufficiently large to cause Republicans to prefer the liberal version of the preschool policy to the conservative version when they are not exposed to arguments (Δ = 1.26, p < .01).

Once more, however, the addition of arguments that explain who benefits from the policy corrects for this reversal, providing further evidence for H2. When party cues are reversed, arguments counteract the influence of parties by increasing Republicans’ support for the conservative proposal (Δ = 0.95, p < .029) and reducing their support for the liberal proposal (Δ = 0.98, p < .028). (See the difference between the coefficients for the reversed party cue and reversed party cue + argument.) Therefore, the reversed party cue effect is eliminated on both proposals when Republicans learn the substantive details of the policies.

The reason arguments can correct the preference order again lies in respondents’ evaluations of competing arguments (see Supplementary Materials). Republicans consistently rate the conservative argument to be more effective than the liberal argument across traditional and reversed party cue conditions, and therefore the arguments tend to push them toward the conservative position on the issue, reinforcing the traditional cue, and offsetting the reversed cue.

Finally, when only party cues but not arguments are present, the ideological direction of the preschool policy has no impact. There is no difference in support for the liberal and conservative versions of the policy if both versions are endorsed by the Republican Party (Δ = 0.49, p < .18), or if both versions are opposed by the Republican Party (Δ = 0.17, p < .37). But, as predicted by H4, once the implications of the preschool policies are fleshed out with arguments, the ideology of the policy has a significant effect on preferences. Support for the conservative version of the elaborated policy is significantly greater than support for the liberal version when the Republican Party opposes both versions (Δ = 1.23, p < .008); respondents also prefer the conservative version of the elaborated policy when the Republican Party endorses both policies, although the difference in support in this case is not statistically significant (Δ = 0.36, p < .25).

Unemployment benefits

Table 3 reveals that Republican respondents’ preferences are clear and stable across experimental conditions on the unemployment issue. This issue is black and white (or more appropriately red and blue) for Republicans as well as Democrats. As we would expect from H3, the direction of the policy makes a difference in the baseline condition (Δ = 0.66, p < .085). Furthermore, the effect of changing the policy direction is significant across conditions, regardless of party cues or arguments. Party cues do not change preferences greatly, nor do arguments (supporting H1b and H2, respectively), because the ideological implications of the unelaborated policy are clear enough to respondents. Indeed, only the traditional party cue on the liberal proposal prompts a significant shift in preferences (β = 0.909).

Summary of Results

Tables A3 to A5 in the Appendix summarize the relative impact of party cues, policy arguments, and policy direction for each issue among Democratic and Republican respondents.

First, we see in Table A3 the effect of party cues is far from constant. For Democrats, it is largest on the tax issue and smallest on the preschool and unemployment issues. The average effect of switching party cues—in the absence of arguments—on both the liberal and conservative proposals of each issue is .77 on the tax issue, .33 on the preschool issue, and a mere .16 on the unemployment issue. Democrats therefore move, on average, about three-fourths of a point on a 7-point scale toward the position endorsed by their party on the liberal and conservative versions of the tax policy. They move less than half this amount on the preschool issue, and barely budge in response to party cues on the unemployment issue.

For Republicans, party is a more powerful cue on the preschool issue, with an average shift in opinion of 1.10 in response to the party endorsements in the absence of arguments. The party cues are slightly less influential on the sales tax issue, moving opinion an average of .81. There is no net shift in opinion on the unemployment issue as the party cues are even less effective among Republicans than among Democrats on this issue.

The potency of arguments follows a similar pattern across issues consistent with our hypothesis that party cues and arguments will have the greatest latitude to move opinion on the least ideologically structured issues. Arguments wield a sizable .80 point average impact among Democrats on the tax issue—the issue they had the most difficulty classifying ideologically, but the corresponding average argument effects for the preschool and unemployment policies are only .06 and .05, respectively.

By comparison, arguments are ineffective on the tax policy among Republicans, who saw this policy in clearer ideological terms; on average, they move opinions slightly in the wrong direction by .07 points. They are even less effective on the unemployment policy, shifting preferences .10 points in the wrong direction. Only on the preschool policy, which Republicans had trouble classifying ideologically, do arguments make a substantial difference, with an average effect of .48 points. This average masks a far more powerful effect of arguments—approaching an entire point on the 7-point scale—when the party cues are reversed on both the liberal and conservative versions of the preschool policy. As noted above, Republicans who are not given policy information through arguments are misled by the reversed party cues to prefer the liberal preschool policy to the conservative version, but they correct the order of their preferences when they are given more detailed information about the implications of the policies.

The conditional effect of party when policy arguments are present is also summarized in Table A3. The party cue has almost the same impact (.72 and .76, respectively) on Democrats’ tax policy preferences whether or not the policies are elaborated by arguments (although when arguments are available on this issue, they exert an independent influence that entirely offsets the influence of party in the reverse cues condition). On the other issue we highlight—the preschool issue among Republicans—the provision of policy information severely diminishes the impact of party cues from 1.10 to only .12 and eliminates the impact of reversed party cues.

Finally, we summarize in Table A5 the effect of changing the policy direction, holding party endorsements constant. This effect is measured both when policies are not elaborated with arguments, and when arguments are provided. The two most striking observations are the negligible effects of the ideological direction of the policy on the tax issue for Democrats and on the preschool issue for Republicans when respondents are not exposed to arguments. Once substantive arguments are provided, Democratic and Republican respondents discern the ideological structure of the issues and express a clear preference for the version of the policy that matches their values. On the tax issue, the average effect among Democrats of changing policy direction is −0.10 (the negative sign indicates Democrats favor the wrong [i.e., conservative] policy) without policy information and 1.50 with information. Likewise, the effect of changing policy direction is negligible among Republicans on the preschool issue when there is no substantive elaboration of the policy (–0.16, again in the wrong direction); but there is a large directional effect (0.79) once the implications of the policy are spelled out.

On the other policy issues, changing the ideological direction of the policy generally has a substantial impact on preferences, even when the policies are not elaborated with information about their implications. The direction of a policy moves preferences among Democrats by more than 2.0 points on the preschool and unemployment issues. Republican preferences change somewhat less on the unemployment issue, but still average well over a full point. In contrast to Democrats, the policy direction on the tax issue changes Republican preferences by 1.20 points even when they have no exposure to arguments, and 1.05 points when policy arguments are supplied.

All three factors examined in this experiment—party cues, substantive policy arguments, and policy direction—can be highly influential under certain circumstances. Both party cues and arguments shape preferences on policies that are not ideologically transparent. But when the ramifications of a policy are more clearly understood, party cues and arguments diminish in value because they cannot significantly change people’s conceptions of the issue. On the more opaque issues where party cues and arguments were influential, their effects were of comparable magnitude. The size of the party and argument effects, however, never equaled the influence of reversing the ideological direction of the policy. Changing the direction of a highly structured policy, such as increasing or decreasing unemployment benefits, produced the largest effects on preferences that we observed in the experiment, sometimes averaging well over 2.0 points.

It is also important to understand how the effects of party, arguments, and policy direction are interrelated. Arguments magnify the consequences of switching the direction of the policy. Effective arguments polarize evaluations of the policy alternatives (liberal vs. conservative versions) and, in so doing, limit the power of reversed party cues to reverse preferences. The more powerful the arguments, the greater the effect of changing the direction of the policy from liberal to conservative, while holding the party endorsement constant. On the tax issue, the policy information contained in the arguments alerts Democrats to the differences between alternative versions of the policy and reduces their reliance on party cues. A similar dynamic occurs on the preschool issue among Republicans, as exposure to arguments substantially checks the influence of misleading party cues and increases responsiveness to the ideological direction of the policy.

Although we have not focused on individual differences, we report variations in mean policy support by experimental condition and political knowledge in the Supplementary Materials. Politically knowledgeable Democrats are somewhat better able to make use of the arguments than are Democrats who are less knowledgeable. Among politically knowledgeable Democrats, the political arguments generally increase support for the liberal versions of the policies and reduce support for the conservative policies. The same pattern of directional movement is also evident, but less pronounced, for the less politically knowledgeable Democrats. We do not observe a consistent pattern of effects of political knowledge among Republicans, but the statistical power of these tests is limited by the smaller number of Republicans in our sample.

Discussion and Conclusion

The results of our experiment provide support for our hypotheses about the relative impact of party cues and policy arguments on policy preferences. By varying issues and manipulating the types of information provided to respondents across three issues, we observed how the ideological clarity of the proposal constrains the influence of party cues. The party cue effect is maximized when the policy is least defined. Substantive arguments also have the greatest potential in this circumstance to shape preferences. This generally fits with our theoretical expectation that both party cues and arguments should exert greater influence when the policy is more ideologically ambiguous.

As in all experimental research, there is a need to consider the potential consequences of pretreatment effects. If party cues fail to move opinion on ideologically structured issues, this does not preclude the possible pretreatment influence of partisanship in the development of opinions on the issue prior to the experiment (Levendusky, 2009; Slothuus, 2016). In our experiment, for example, the tax issue was amorphous in the control conditions, leaving room for party cues and arguments to move opinions. If this issue were to make the policy agenda, a sustained partisan discussion of its implications could produce a more structured issue that clearly divided along ideological lines, in the manner induced by our experimental treatments. As a result, we would expect that party cues and policy details would have less latitude to shape preferences subsequently because people already responded to those cues in the pretreatment period. Indeed, party cues may have had less impact on the unemployment issue because the political parties have consistently divided on this issue along ideological lines, making it easier for people to evaluate this policy without party cues. As Slothuus (2016) noted, an experiment may “find no effect of parties at the very time that their influence is strongest outside the experiment” (302). However, an equally important corollary, substantiated in our experiment, is that political parties and elites should be similarly constrained in their ability to reverse stances on issues that the public has sorted on.

Adding information in the form of arguments can reduce uncertainty over a policy and, in so doing, restrict the latitude of party cues to move preferences. Arguments magnify the contrast between liberal and conservative versions of a policy when the ideological direction of the policy is hard to discern in the baseline condition. If the consequences of the policy are easily understood in the baseline condition, then further information is superfluous. But when policies are initially ambiguous, substantive information conveyed through arguments can prove invaluable by increasing the contrast between policy alternatives. The magnitude of the effect of changing the policy direction corresponds to our predictions for the three issues we studied. Democrats’ preferences on the tax issue, and Republicans’ preferences toward the preschool policy, were least affected by changing the policy direction until more policy information was provided. In contrast, both Democrats’ and Repubicans’ preferences on the unemployment issue experienced the most dramatic shift when the policy direction changed, with or without additional information.

If preferences depend on both party cues and arguments, both forces will work in tandem in the traditional cues condition (pushing apart preferences toward the liberal and conservative policies), but will be in conflict in the reversed party cues condition. Only when arguments are perceived consistently across conditions can they counteract reversed party cues (see Supplementary Materials). The best illustration of this point is the contrast between Democratic and Republican responses to the tax issue. Republicans did not consistently evaluate the conservative policy arguments to be stronger than the liberal arguments across traditional and reversed party conditions, so these arguments provided only weak resistance to the reversed party cues. Conversely, Democrats consistently evaluated the liberal arguments to be stronger than the conservative arguments across conditions; therefore, when the policies were embellished with arguments, Democrats held a correct preference ordering of the competing policies even when party cues were misleading (i.e., reversed).

Overall, the findings from our experiment give us reason to be mildly optimistic: people can take and maintain positions on issues despite shifting party cues. A further encouraging sign is that people are responsive in sensible ways to the substantive details of information. Respondents updated their interpretations and evaluations of policies in accord with the information they received about the policies’ social and economic consequences.

Consistent with past research on mass belief systems (Converse, 1964), we expect some individuals will be better able to recognize the ideological structure of issues and to exhibit ideological consistency in their policy preferences. More recent studies have found that ideological consistency across issues has increased over time in today’s polarized political environment largely among the segment of the public that is most attuned to elite cues (Jewitt & Goren, 2016; Lupton, Myers, & Thornton, 2015). As noted above, we find some evidence that politically sophisticated Democrats made somewhat greater use of the ideological information contained in the policy arguments than did those who were less knowledgeable. Any increase in the electorate’s ideological constraint across policy attitudes and in the consistency of their ideological identifications and policy attitudes ought to reduce the latitude parties have to adopt ideologically inconsistent positions without repercussions (Bawn, Cohen, Karol, & Masket, 2012; Lauderdale, Hanretty, & Vivyan, 2018).

Our theoretical focus on the clarity of ideological divisions on an issue is distinct from previous research emphasizing issue salience and the personal importance of issues (e.g., Ciuk & Yost, 2016). We would hypothesize that even on a personally important issue—such as the tax issue in our experiment—parties can enjoy considerable scope in adopting either side of the issue if the alternatives are not structured in conventional ideological terms. Conversely, we would predict, contrary to the salience hypothesis, that preferences on an issue of low personal relevance—such as a proposal to change unemployment benefits evaluated by individuals who are economically secure—would nonetheless be less susceptible to party cues if the policy could be readily classified as either liberal or conservative.

Our study also has implications for discussions of possible asymmetries between Democratic and Republican identifiers (e.g., Grossmann & Hopkins, 2015) in their responsiveness to party cues and policy details (Bullock, 2011). To the extent Democrats were at times more (or less) responsive than Republicans to cues and arguments, this can probably be explained by the characteristics of the issues chosen for study rather than a consistent difference between partisans across issues. Partisan differences in our study resulted from the relative ease or difficulty that Democrats and Republicans had in classifying our three issues. In the baseline condition, Democrats found the tax issue to be more ideologically ambiguous than did Republicans, while Republicans had similar difficulty classifying the preschool spending policy. Because of their ambivalence toward the respective policies, Democrats were more responsive to party cues and arguments on the tax issue than were Republicans, but the opposite was true on the preschool issue, where Republicans may have exhibited the kind of conflicted conservatism documented by Ellis and Stimson (2012). Such contrasting results owing to the details of the policies should make scholars cautious about drawing broad inferences about partisan differences from single-issue studies.

In general, the additional information people receive about a policy changes their preferences if it updates their understanding of the relationship between the policy and their political values. Boudreau and Mackenzie (2014) similarly showed that providing information about ballot initiatives could move public opinion, but our findings regarding the effects of information are stronger. In their study, information moved preferences only when it was one-sided, but when information was balanced, people relied on party cues. We found instead that competing arguments presented simultaneously can influence respondents’ preferences and reduce the effect of party cues. Most important, policy information that identified the intended effects and beneficiaries of a policy influenced our respondents even when their own party adopted and defended a contrary position.

Supplemental Material

Supplementary_Materials_7-12-18_(1) – Supplemental material for Information and Issue Constraints on Party Cues

Supplemental material, Supplementary_Materials_7-12-18_(1) for Information and Issue Constraints on Party Cues by Dennis Chong and Kevin J. Mullinix in American Politics Research

Footnotes

Appendix

	Republicans’ change in preference
	No arguments			Arguments
	Republicans support	Republicans oppose	Issue average	Republicans support	Republicans oppose	Issue average
Tax	0.726	1.662***	1.194	1.045**	1.058**	1.052
Preschool	−0.494	0.170	−0.162	0.36	1.227***	0.794
Unemployment	1.107**	1.815***	1.461	1.321***	1.217**	1.269

Note. A negative coefficient indicates a preference change that is contrary to the expected effect of changing the policy direction. Significance tests not conducted for issue average.

p < .05, one-tailed. ***p < .01, one-tailed.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Supplemental Material

Supplemental material for this article is available online.

Notes

Author Biographies

Dennis Chong is professor of Political Science at the University of Southern California and the John D. and Catherine T. MacArthur Professor of Political Science (ret.) at Northwestern University. He studies individual and collective decision-making in politics and society.

Kevin J. Mullinix is an assistant professor in the Department of Political Science at the University of Kansas. His research concentrates on public opinion and public policy.

References

Arceneaux

(2008). Can partisan cues diminish democratic accountability? Political Behavior, 30, 139-160.

Arceneaux

(2012). Cognitive biases and the strength of political arguments. American Journal of Political Science, 56, 271-285.

Bawn

Cohen

Karol

Masket

(2012). A theory of political parties: groups, policy demands, and nominations in American politics. Perspectives on Politics, 10, 571-597.

Bolsen

Druckman

J. N.

Cook

F. L.

(2014). The influence of partisan motivated reasoning on public opinion. Political Behavior, 36, 235-252.

Boudreau

MacKenzie

S. A.

(2014). Informing the electorate? How party cues and policy information affect public opinion about initiatives. American Journal of Political Science, 58, 48-62.

Bullock

J. G.

(2011). Elite influence on public opinion in an informed electorate. American Political Science Review, 105, 496-515.

Bullock

J. G.

Gerber

A. S.

Hill

Huber

G. A.

(2015). Partisan bias in factual beliefs about politics. Quarterly Journal of Political Science, 10, 519-578.

Campbell

Converse

P. E.

Miller

W. E.

Stokes

D. E.

(1960). The American voter. New York, NY: Wiley.

Chong

Druckman

J. N.

(2007). Framing public opinion in competitive democracies. American Political Science Review, 101, 637-655.

10.

Ciuk

D. J.

Yost

(2016). The effects of issue salience, elite influence, and public opinion in an informed electorate. Political Communication, 33, 328-345.

11.

Clifford

Jewell

R. M.

Waggoner

P. D.

(2015). Are samples drawn from Mechanical Turk valid for research on political ideology? Research and Politics, 2(4), 1-9.

12.

Cohen

G. L.

(2003). Party over policy: The dominating impact of group influence on political beliefs. Journal of Personality and Social Psychology, 85, 808-822.

13.

Converse

P. E.

(1964). The nature of belief systems in mass publics. In Apter

(Ed.), Ideology and discontent (pp. 206-261). New York, NY: Free Press.

14.

Coppock

(2018). Generalizing from survey experiments conducted on Mechanical Turk: A replication approach. Political Science Research and Methods, 1-16. doi:10.1017/psrm.2018.10.

15.

Delli Carpini

Keeter

. (1996). What Americans know about politics and why it matters. New Haven, CT: Yale University Press.

16.

Druckman

J. N.

(2001). Using credible advice to overcome framing effects. Journal of Law, Economics, and Organization, 17, 62-82.

17.

Druckman

J. N.

Peterson

Slothuus

(2013). How elite partisan polarization affects public opinion formation. American Political Science Review, 107, 57-79.

18.

Ellis

Stimson

(2012). Ideology in America. Cambridge, UK: Cambridge University Press.

19.

Grossmann

Hopkins

D. A.

(2015). Ideological republicans and group interest democrats: The asymmetry of American party politics. Perspectives on Politics, 13, 119-139.

20.

Iyengar

Westwood

S. J.

(2015). Fear and loathing across party lines. American Journal of Political Science, 59, 690-707.

21.

Jewitt

C. E.

Goren

(2016). Ideological structure and consistency in the age of polarization. American Politics Research, 44, 81-105.

22.

Kam

(2005). Who toes the party line? Political Behavior, 27, 163-182.

23.

Lauderdale

Hanretty

Vivyan

(2018). Decomposing public opinion variation into ideology, idiosyncrasy, and instability. Journal of Politics, 80, 707-712.

24.

Levendusky

M. S.

(2009). The partisan sort: How liberals became democrats and conservatives became republicans. Chicago, IL: University of Chicago Press.

25.

Levendusky

M. S.

(2010). Clearer cues, more consistent voters. Political Behavior, 32, 111-131.

26.

Lupton

R. N.

Myers

W. M.

Thornton

J. R.

(2015). Political sophistication and the dimensionality of elite and mass attitudes, 1980-2004. Journal of Politics, 77, 368-380.

27.

Malhotra

Kuo

A. G.

(2008). Attributing blame: The public’s response to Hurricane Katrina. Journal of Politics, 70, 120-135.

28.

Mullinix

K. J.

(2016). Partisanship and preference formation. Political Behavior, 38, 383-411.

29.

Mullinix

K. J.

(2018). Civic duty and political preference formation. Political Research Quarterly, 71, 199-214.

30.

Mullinix

K. J.

Freese

Druckman

J. N.

(2015). When convenience samples yield generalizable estimates in social science experiments. Journal of Experimental Political Science, 2, 109-138.

31.

Nicholson

S. P.

Hansford

T. G.

(2014). Partisans in robes: Party cues and public acceptance of supreme court decisions. American Journal of Political Science, 58, 620-636.

32.

Popkin

S. L.

(1991). The reasoning voter: Communication and persuasion in presidential campaigns. Chicago, IL: University of Chicago Press.

33.

Rahn

W. M.

(1993). The role of partisan stereotypes in information processing about political candidates. American Journal of Political Science, 37, 472-496.

34.

Redlawsk

D. P.

Civettini

A. J. W.

Emmerson

K. M.

(2010). The affective tipping point. Political Psychology, 31, 563-593.

35.

Riggle

E. D.

Ottati

V. C.

Wyer

R. S.

Kuklinski

Schwarz

(1992). Bases of political judgments. Political Behavior, 14, 67-87.

36.

Robison

Mullinix

K. J.

(2016). Elite polarization and public opinion. Political Communication, 33, 261-282.

37.

Slothuus

(2016). Assessing the influence of political parties on public opinion. Political Communication, 33, 302-327.

38.

Taber

C. S.

Cann

Kucsova

(2009). The motivated processing of political arguments. Political Behavior, 31, 137-155.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.87 MB