Challenging by Cueing? An Investigation of Party and Leader Cueing Effects Across Mainstream and Challenger Party Voters
- Open Access
- 17.07.2024
- Original Paper
Erschienen in:
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by (Link öffnet in neuem Fenster)
Abstract
Much of the existing work on preference formation has focused on the role of political parties in influencing policy opinions (Campbell et al., 1960; Green et al., 2002; Leeper & Slothuus, 2014). According to this literature, citizens use policy endorsements from the parties they support as cues to position themselves on political issues, converging to that of their preferred party (Brader & Tucker, 2012; Brader et al., 2012; Bullock, 2011; Coan et al., 2008; Kam, 2005; Merolla et al., 2008; Pannico, 2017; Petersen et al., 2010, 2012; Slothuus & Bisgaard, 2020; Slothuus & de Vreese, 2010; Torcal et al., 2018). One assumption is that partisan cues would be more effective the more familiar individuals are with parties, since time is key for the development of party attachments – regardless of whether these are the result of information updating or psychological identification.
While this may be the case for mainstream parties, the effect of cues may not rely on these factors for new challenger parties. Due to their newness, citizens do not have much familiarity with challenger parties, nor do they know how they would behave in government. This does not imply that cues from challengers should be less effective. Instead, they might be conveyed by other sources, or their strength might be determined by different conditions. In this regard, challenger parties often rely on personalistic leadership styles (Garzia et al., 2022), which may make cues from their leaders more effective than those from the party. Moreover, challengers tend to focus on cultural and emotionally charged issues, profiting from an increase in affective politics (Gidron et al., 2022; Iyengar et al., 2012). This focus may allow them to develop social identities ties along with an expressive component of partisanship (Huddy & Bankert, 2022). Alternatively, mainstream parties may benefit from their established brand values. As a result, their cueing effects may be primarily boosted by instrumental partisanship and evaluations of party competence reputation (Achen, 2002; Downs, 1957; Fiorina, 1981).
Anzeige
This article aims to reassess the party-model of opinion formation by examining the role of partisan and leader cues among voters of mainstream and challenger parties. Moreover, it analyzes two conditional factors that can enhance cueing effectiveness: the strength of partisan social identities1 and party competence evaluations. This approach allows for a more definitive test of how cueing effects vary across the two types of party supporters. We achieve this by analyzing data from a cueing experiment involving two issues with varying degrees of complexity, embedded in an online survey conducted in Spain in 2015. At that time, Spain was characterized by a mix of mainstream and challenger parties spanning the ideological spectrum. In the following sections, we demonstrate the presence of a consistent partisan cueing effect that is similar across types of cues and, importantly, across voters from different parties. While the results are more nuanced for certain policy issues, there are indications that cueing effects for voters of mainstream parties tend to blend expressive and instrumental reasoning. In contrast, cueing effects for voters of new challenger parties are somewhat conditioned by the perceived party competence reputation, suggesting the importance of instrumental reasoning in this group of supporters.
This article contributes to the literature on opinion formation regarding cueing effects (Bolsen et al., 2013; Bullock, 2020) and the nature of partisanship in contemporary democracies (Huddy & Bankert, 2022). Firstly, it demonstrates that cueing effects are relevant to voters of different party types, suggesting that new challenger parties might be as effective as mainstream parties in influencing preferences. This finding questions the assumption that cueing effects are enhanced by familiarity and time. Secondly, it indicates that, despite the emphasis on leadership in the personalization literature, leader cues are not more significant than party cues, even for new challenger parties. Furthermore, the study investigates the conditions under which cueing by challenger parties is more effective. It suggests that the potential of new challenger parties and their leaders to induce changes and influence public opinion is mainly contingent upon pre-existing assessments of party reputation rather than social identification. These findings have implications for the ongoing debate regarding the increasing relevance of affective politics and the view of partisanship as primarily an emotional attachment through partisan social identification, which motivated reasoning approaches connect to how party cues operate. However, the evidence also highlights the need for a more comprehensive perspective on how partisan cueing encompasses both expressive and instrumental types of party attachments to varying degrees depending on the type of party.
Theoretical Framework
The Influence of Political Parties on Public Opinion Formation
In representative democracies political parties are one of the most influential forces in shaping citizens’ political attitudes and behaviors, as they help structure political competition between groups by mobilizing public opinion and collective action regarding relevant issues (Aldrich, 1995). Two approaches, found at the intersection between theories of bounded rationality (Lupia et al., 2000), dual-process models (Bullock, 2020) and motivated reasoning (Kunda, 1990; Taber & Lodge, 2006), help explain how parties may influence public opinion.
The first approach, more in line with dual-process models and party cues as informational shortcuts, starts from the premise that the electorate often lacks basic political knowledge, as citizens tend to avoid the costs of seeking and processing information (Kuklinski & Quirk, 2000; Sniderman et al., 1991). However, low levels of knowledge do not prevent citizens from making choices in line with their own policy preferences. By publicly stating their positions, political parties provide a simplified menu of choices across a broad range of policy areas so that citizens can coordinate their orientations with those of likeminded parties to find consistency in public opinion. According to this informational logic, citizens would use party endorsements as cues to position themselves on issues, using simple decision rules rather than a detailed examination of the content of the policy at stake (e.g., Lau & Redlawsk, 2006; Mondak, 1993; Sniderman, 2000). In this way, heuristic processes are a useful tool to achieve more accurate and competent decisions by choosing the policy option with the greatest proximity to the voter’s ideal position, coming to a similar decision as if one had full knowledge on the topic (Hobolt, 2007; Kuklinski & Quirk, 2001).
Anzeige
The second approach, which finds in motivated reasoning the most relevant proponent, looks at the ability of parties to create a strong source of identity and emotional attachment, guiding citizens in their interpretation of political information and their political choices (e.g., Lavine et al., 2012; Muirhead & Rosenblum, 2020; Sniderman, 2000; Taber & Lodge, 2006). Following this view, when a political party advocates a certain position, partisan leanings are activated, so that citizens will tend to align their position to that of the party they feel close to or support in electoral competition. Thus, party labels are used as means to achieve directional goals, or to conform to group norms by adjusting one’s policy positions to fit with their partisan attachments.
These two approaches assume that partisan attachments should be the key conditioning factor for the effectiveness of party cues. The concept of partisanship, however, has generated a long-standing debate which has crystalized around partisanship as either expressive or instrumental (Huddy & Bankert, 2022). The original concept of partisanship dates back to the “Michigan School” (Campbell et al., 1960), which views partisan identification as a product of early socialization resulting in a profound psychological bond that shows great stability and resistance to political and economic change. Further revisions of this model complement this view emphasizing instrumental interpretation of partisan attachments. According to this amendment, the linkage between parties and citizens could also be the result of performance evaluations. Thus, partisan attachment would be continuously updated with exposure to new information and an individuals’ calculations as to the benefit of having a certain party in power would be the driving force behind party ties, instead of ingroup sentiments (Achen, 2002; Downs, 1957; Fiorina, 1981).
Last, a social psychology approach to party attachments has developed an alternative model of expressive partisanship based on the idea that parties connect social groups with individual values and expectations, resulting in some form of a partisan social identity with a substantial motivational component (Huddy, 2015; Huddy, 2018). Those who perceive themselves as group members are motivated to protect their party status feel the desire to positively distinguish the group from others and develop of ingroup bias (Tajfel & Turner, 1979), fueling motivated reasoning and partisan sorting, encouraging the alignment between party identity and issue attitudes (Mason, 2018). In both interpretations, experience and familiarity with parties are fundamental to develop partisanship, as people acquire new information or develop identification with groups only thanks to the passage of time which should make partisan attachments and cueing effects stronger.
Party Cueing and New Challenger Parties
Party cueing effect refers to the extent to which policy opinions of citizens can be influenced by their exposure to the point of view of a political party, regardless of the content of the position or its framing. One efficient solution to identify party cueing has been to manipulate the cues along with policy issues in an experimental setting. Experimental studies have shown significant effects of partisan cueing on individuals’ positions in the United States (Arceneaux, 2008; Bullock, 2011; Coan et al., 2008; Kam, 2005; Nicholson, 2012). Research has demonstrated that party cueing is also found, with differences in the size of the effect, in more fragmented multi-party systems where social and political cleavages other than partisanship may be more important influences on citizens’ opinions (Torcal et al., 2018). These findings have been consistent for traditional democracies with stable party systems (Merolla et al., 2008; Petersen et al., 2012; Slothuus & De Vreese, 2010; Sniderman & Hagendoorn, 2007), as well as in younger democracies with less institutionalized party systems, though in these countries cueing effects have been found weaker (Brader & Tucker, 2012; Brader et al., 2012; Conroy-Krutz, 2016; Merolla et al., 2007; Samuels & Zucco, 2014).
Scholars have also focused on the conditioning role of party age on cueing with mixed results, showing that age moderates cueing effects but that this role is strengthened or reduced when other factors pertaining to the ideological platform or incumbency are accounted for (Brader et al., 2012). Recent research also indicates that cueing effects apply to new parties (Brader et al., 2020; Pannico & Anduiza, 2020). Yet, these studies overlook the role of different source cues and how alternative components of partisanship may be at work across different types of voters.
A more definitive test ought to consider the evolving nature of challenger parties and its potential connection with the increasing “personalization” of politics, defined as the gradual shift away from electoral decision-making based on social groups and the relative increase in the significance of individual leaders (Garzia et al., 2022). This is especially relevant due to the increasing phenomenon of party-voter dealignment, the weakening of party organizations, and the advancements in new communication technologies. Additionally, with the shift towards media-centered campaigns, leaders have gained the ability to promote themselves more independently from the party they represent (Aarts et al., 2011; Costa Lobo & Curtice, 2015; Karvonen, 2010).
This is particularly pertinent for challenger parties (De Vries & Hobolt, 2020). Leaders of this type of party play a central role in connecting voters by transmitting their innovative positions on the different issues and, as such, leaders should be considered a more significant source cue when analyzing the partisan cueing effects across mainstream and challenger political formations. In brief, it is reasonable to expect cues sent by party organizations to be more effective among mainstream voters while those issued by leaders to have a greater effect for challenger voters. Hence, our first two hypotheses posit:
H1
The effect of party cues on citizens’ policy positions will be stronger for voters of mainstream parties than for challenger parties.
H2
The effect of leader cues on citizens’ policy positions will be stronger for voters of challenger parties than for mainstream parties.
Another missing aspect in recent comparative literature on party cueing is the increasing presence of affective politics in contemporary democracies (Ipsos, 2018). Connected to this phenomenon, challenger parties aim at counteracting the advantaged position of mainstream parties by adopting more salient and extreme positions on cultural disputes that appeal to basic emotions (Gidron et al., 2022) promoting stronger partisan social identities among their supporters (Huddy & Bankert, 2022). The affective component of these social ties may increase the distinctiveness of cues, and, in turn, the possibility of attitudes that are more consistent to their party preferences (Mason, 2018). Thus, affective politics could be increasing the capacity of partisan social identities to influence public opinion, especially for challenger party voters. Consequently, when examining the presence of partisan and leader cueing effects among voters, it is reasonable to anticipate that these effects could be influenced by partisan affinities. Thus, our third hypothesis anticipates that:
H3
The moderating role of in-party group affinities on cueing will be stronger among voters of challenger parties compared to those of mainstream parties.
On the other hand, the defining feature of mainstream parties is their dominance within the political market regardless of the party family to which they belong (De Vries & Hobolt, 2020). Thus, such organizations have accumulated government experience controlling the provision of policies and have had the opportunity to become familiar to voters over a longer period. Thus, mainstream parties derive an advantage from possessing a well-established track of information (Achen, 2002), which becomes an integral component of their reputation (Coan et al., 2008). This should put them in a more favorable position for cueing opinions as they may rely on more recognizable established brands, and party competence reputation should be more important in moderating the role of cues as these perceptions should be rooted in actual policy experience. Conversely, the absence of policy and performance records might be one of the impediments that new challenger parties face and would likely make their cues less dependent on instrumental considerations related to party competence reputation. Hence, our last hypothesis anticipates that:
H4
The moderating role the evaluation of in-party competence reputation on cueing will be stronger among voters of mainstream parties compared to those of challenger parties.
Data and Methods
We evaluate the effect of partisan cues using an experiment included in the ‘CIUPANEL’ dataset (Torcal et al., 2016) conducted on a sample of the Spanish population in the ten days before the general elections held in December 2015. We assess the relative role of party and leader cues across different parties in a between-subject design, allowing for stricter control and reducing the role of confounding variables via randomization of treatments. We test the relevance of various cues on policy issues with varying complexity and resembling existent positions to mimic real world choices in a very diverse party system.
The Spanish Context
Spain is a useful case as it provides a context for a robust test of partisan cueing effects across older mainstream and newer challenger parties. The country has experienced one-party government with alternation of power among the two traditional parties – the Spanish Socialist Party (PSOE) and the center-right People’s Party (PP) – together with stable voter orientations from its transition at the end of the 1970s until 2011. However, from that year onward, Spain has struggled with increased fragmentation and a profound process of party system restructuring, allowing for the emergence of new challenger political parties. Podemos (P’s), for example, unexpectedly obtained 8% of the vote share and five seats in the European Parliament in the 2014 European election. Podemos remains one of the main actors in the Spanish party system and in the 2015 general election (jointly with its regional partner organizations) it gained about 21% of the vote. In the same election, a second new challenger party, Ciudadanos (C’s), also obtained representation in the national parliament for the first time with 14% of the vote share. The visibility of these parties might be connected to the personalization of politics in that popularity of their leaders was important to their success (Orriols & Cordero, 2016). Overall, the co-existence of challenger and traditional parties in the same system makes Spain one of the few cases in which we can rigorously test different cueing effects across political formations.
Experimental Design
Our survey experiment follows preceding studies (e.g., Brader et al., 2020), in that we randomly assigned subjects to one of three treatment groups. In the control group (no cueing), the selected respondents were asked to choose one of several policy options but without showing any informative cue. In the party cueing treatment group, the same policy positions were labeled with the endorsement of a specific party. In the leader cueing treatment group, the same policy positions were labeled but with the party leader endorsing a specific option. Balance tests give positive indication of the robustness of random assignment, as reported in the online appendix (Tables A1-A3).
The two issues selected and received by the three treatment groups regarded the ‘Reform of the Senate,’ and ‘Health assistance and insurance to irregular migrants’. To avoid priming effects, the issues were presented in a randomized order. We intentionally chose two issues with relatively low levels of saliency, but with different levels of complexity (Carmines & Stimson, 1980). The ‘Reform of the Senate’ was a very technical issue related with the model of decentralization of the State and required high levels of informational content. The issue of ‘Health assistance for immigrants’ was not highly salient at the time of the study, but it is a topic with a basic cultural content that has sparked affective divisions among different partisan groups (Torcal & Comellas, 2022), requiring less information processing compared to more complex issues.
Figure 1 displays the experimental stimuli. As can be seen, positions on these issues range from maintaining the status quo at that time to more critical stances. The descriptive statistics related to the experimental questions can be found in the online appendix (Table A4). The policy issues and the positions taken by leaders and parties in the cues provided closely resembled existing policy debates and positions at the time of the survey’s implementation in an attempt to increase external validity and diminish respondent deception.
Fig. 1
Experimental stimuli for the 2015 General elections
We included a broad set of political formations and their main leaders and matched them with the issue position they supported at the time of the electoral campaign. The following mainstream parties and their leaders were considered: the center-left Partido Socialista Obrero Español (PSOE) and its leader Pedro Sánchez, and the center-right to right-wing Partido Popular (PP) and its leader Mariano Rajoy. We also included new challenger parties such as the centrist Ciudadanos (Cs) and its leader Albert Rivera and the leftist Podemos and its leader Pablo Iglesias. Last, we included small regional leaders and parties, such as the Catalan regional nationalist parties – Convergència Democràtica de Catalunya (CDC, center-right) and Artur Mas, – and the Basque nationalist party – the Partido Nacionalista Vasco (PNV, center-right) and Iñigo Urkullu, although those are not the focus of the current study.
This design presents several advantages. First, a broad set of parties and leaders from different political formations over the whole political spectrum are included. Additionally, by incorporating issues with different levels of complexity, we allow for a stronger test of our hypotheses. Last, building on Nicholson (2012), different source cues enable a more exhaustive assessment of two types of partisan cueing among different types of parties.
Variables and Models
Our dependent variable measures whether the policy positions chosen by citizens in our experiments are congruent with their relative party preferences (e.g., Brader et al., 2020; Pannico, 2017; Torcal et al., 2018). Thus, we have created a dichotomous dependent variable which takes the value of ‘1’ when the respondent chooses the option endorsed by her or his preferred party or leader – that is, if the conditions for congruence are met – and ‘0’ otherwise. We identify respondents’ party preferences by using vote probability (van der Eijk et al., 2006) in the pre-electoral survey preceding the 2015 general election.2 As such, in our analysis we focus on respondents who expressed a preference for one of the six considered parties. To evaluate the effect of partisan cues, we estimate a series of logistic regression models using the following baseline equation:
$$Pr \left({y}_{i}=1\right) ={{logit}^{-1}(\beta }_{0}+{\beta }_{1}{partycues}_{i}+{\beta }_{2}{leadercues}_{i}).$$
(1)
where we model the probability of success \(\left(y=1\right)\) in each observation \(i\), namely whether the voter expresses a position which is congruent with the preferred party. This is linked to the intercept, \({\beta }_{0}\), and beta-coefficients \({\beta }_{1}\) and \({\beta }_{2}\) of other covariates, gauging the effect of the different treatments. The covariates are dummy variables resulting from a categorical measure which identifies whether the respondent is assigned to the control group (the reference category) or to a party cue (policy endorsed by a party) or to a leader cue (policy endorsed by a leader). To assess whether the effect of both cues varies across types of party voters we run separate models for each subsample derived by the single party preference, and for party preferences grouped according to either old mainstream (PP and PSOE) or new challenger options (Podemos and Ciudadanos).
To test our last two hypotheses, we let cue treatments interact either with a measure of individual affinity towards one’s own group of voters and, later, with an indicator of one’s own party competence reputation, respectively. We present separate models across voters focusing again on old mainstream (PP and PSOE) and new challenger options (Podemos and Ciudadanos) to disentangle whether cueing hinge on different conditions. We measure individual in-group affinities on a 0 to 10 (unfavorable-favorable) feelings scale towards their respective own party voters (we match the score each respondent assigns to the group of belonging, if this is the highest across a list of groups of voters). Positive affect towards one’s favored voter group is a useful indicator of expressive partisanship understood as the manifestation of social identities (Huddy et al., 2020; Rosema et al., 2020) and it is considered a valid measure of in-group party affinity (Tichelbaecker et al., 2023) able to predict political behavior (Iyengar et al., 2019; Wagner, 2021).
When it comes to parties’ competence reputations, we use respondents’ evaluations of which party was better able to handle five more general issues (unemployment, education, health, immigration and corruption). This type of indicator has been commonly used to assess opinions about party competence or performance (Green et al., 2002). Our final index is an additive score of one’s own party issue credibility. To enhance comparability and deal with distribution skewness we break the two measures3 – in-party group affinity and in-party credibility scales – splitting the sample on each one’s median value.
Since results of interactions are not easy to interpret, especially in non-linear models, we base our conclusions on average marginal effects and tests for differences (Ai & Norton, 2003; Mize et al., 2019). For the sake of simplicity, we display graphical results showing post-estimations and comment on the main findings in the text. For question wording, measurement, as well as for the complete results, see the online appendix (Tables A6-A16).
Sample and Participants
The data comes from one of the waves of an online longitudinal six-wave panel survey (Torcal et al., 2016) conducted during 2014 and 2015 using a non-probabilistic sample of citizens aged 18 or over with access to the Internet either at home or at work. The respondents were recruited by active invitation among registered online panelists. Self-registration was not allowed. Quotas were employed to approximate a sample representative of the general population along key sociodemographic variables (gender, age, size of locality, autonomous community of residence). Although this is not a probabilistic sample, the distribution of important socio-demographic, partisan, and ideological characteristics in our sample approximates the same parameters obtained by the gold standard of the Spanish National Centre for Sociological Research (CIS) during the same years (see the online appendix, Table A5). The only differences are that our sample is skewed slightly towards the more educated than the general population (National Institute of Statistics – INE – Census data 2015). Despite this difference, recent studies have revealed considerable similarities between experimental treatment effects obtained from convenience samples and representative population samples (e.g., Mullinix et al., 2015). We confine our analysis to the fifth wave (pre-election survey in 2015) which includes the experiment relevant for this study.4
Results
Partisan Cue Taking across Types of Party Voters
We start by analyzing the relevance of partisan cues in general, and then we address the question of whether cueing effects depend on the type of party. Figure 2 displays the Average Marginal Effects (AMEs) obtained after regressing our measure of policy congruence for the two issues on the different partisan labels – party and leader cue treatments – under different conditions (models are shown in Tables A6-A11 in the online appendix).
Fig. 2
Party and leader cueing effects. Note Lines across dots represents 95% confidence intervals. Estimates represents Average Marginal Effects of Logistic regression models predicting policy congruence given the assignment of partisan cues relative to the control (the dotted line)
Starting from the top of the two panels, we find post-estimations for the six considered party voters as taken altogether (the “voter sample” estimation). As can be seen, the level of congruence is always higher for respondent groups that received one of the two informational cues. In other words, when clear cues are present, voters tend to use them to align their policy position with the endorsement of either the party or leader they support. Moreover, the effects produced by the examined cues are relevant for the two policy areas. In fact, point estimates for the treatment conditions are always more positive than the control condition (no cue represented by the dotted line). However, the test performed on the effects across models confirms that partisan cueing seems to be stronger for the more complex issue, the ‘Reform of the Senate’, while this is effect is weaker but still significant for the issue concerning health care and assistance for immigrants (Table A10). Overall, we do not encounter significant differences in the effects across the two cues for any of the issues considered. Specifically, for the issue of the ‘Reform of the Senate’, the percentage of respondents choosing the policy position adopted by their preferred party increases by 34% points (SE = 0.031; p < 0.001; N = 1288) and 32% points (SE = 0.030; p < 0.001; N = 1288) when a party or a leader cue is provided, respectively. On the other hand, for the issues concerning the ‘Health care and assistance for immigrants’, the party cueing effect corresponds to an increase of 21% points (SE = 0.034; p < 0.001; N = 1244) in comparison to an increase of 23% points for the leader cueing effect (SE = 0.034; p < 0.001; N = 1244).
These results demonstrate the role of leader cues beyond that of party labels, but the difference in the effects produced by the two is undistinguishable. This evidence complements existing research on party cueing effects (e.g., Brader et al., 2012; Brader et al., 2020; Bullock, 2011; Coan et al., 2008; Kam, 2005; Merolla et al., 2008; Nicholson, 2012; Petersen et al., 2010, 2012; Slothuus & De Vreese, 2010; Sniderman & Hagendoorn, 2007; Samuels & Zucco, 2014; Torcal et al., 2018), while contrasting with the literature on the personalization of politics that posits the greater importance of leaders over parties (e.g., Garzia et al., 2022; Karvonen, 2010), at least as far as leader cueing effects and public opinion formation are concerned.
Centering our attention on our main argument, do we see a difference in cueing effects across voters for old mainstream and new challenger parties? At the bottom of the panels, we display the post-estimation for each one of the considered supporter groups. Results show that cueing effects are significant for all the considered voters, regardless of the type of cue (for ease of reading, for the ‘Reform of the Senate’, we omit here results for nationalist parties which are still significant, despite the small sample size, see Figure A1 in the online appendix). In fact, we do find an increase in policy congruence for all the considered party groups (only leader cues do not seem to have a significant effect for Ciudadanos voters on the issue concerning healthcare provision for immigrants, although the effect is in the expected direction). More importantly, this first result paves the way for testing whether party cueing should be stronger for voters of old mainstream parties (H1), while leaders should be more effective in cueing opinions among supporters of new challenger political formations (H2). If we turn our attention to the two post-estimations at the bottom of the panels, for mainstream voters party cueing effects correspond to an increase of 34% points (p < 0.001; N = 494) for the ‘Reform of the Senate’ issue and 26% points for the ‘Health care and assistance for immigrants’ one (p < 0.001; N = 494); while for challenger voters this is equal to 33% points (p < 0.001; N = 750) and 17% points (p < 0.001; N = 750), respectively. Regarding leader cueing effects, results show that for mainstream voters this corresponds to an increase of 28% points (p < 0.001; N = 494) for the ‘Senate’ issue and 24% points for the ‘Immigrants’ one (p < 0.001; N = 494); while for challenger voters this is equal to 32% points (p < 0.001; N = 750) and 22% points (p < 0.001; N = 750). In light of the tests performed on these effects, we conclude that party and leader cueing effects do not vary in a significant way according to party groups, leading us to reject both hypotheses (H1 and H2): the effect of partisan cues is undistinguishable across types of party voters (Table A11).
It is worth noting that one potential problem in cueing experiments, in particular with real-world issues, might be pre-treatment effects (Slothuus, 2016). Policy positions are likely to be known by respondents, leading to an underestimation of the treatment effects. This might be the case of supporters of older mainstream parties who should be more familiar with the positions of their parties, which could in turn undermine our comparison across types of parties. We test this by analyzing whether policy congruence varies according to the group of voters in the control group. For neither of the two issues are voters of old mainstream parties more likely to choose the position held by their party.5 This suggests that our results and conclusions are reliable (see Figure A2 in the online appendix).
Overall, these findings somehow contradict the empirical research connecting the age of parties with their capacity to convey cues (Brader et al., 2012), while providing further support to studies that maintain new challenger parties are able to cue opinions in spite of less familiarity with the electorate and no incumbent record (e.g., Pannico & Anduiza, 2022; Torcal et al., 2018). The fact that leader and party cueing is not different among challenger voters provide further confirmation that new challenger parties, as political organizations, might be as influential as old mainstream parties. A series of tests showing the robustness of our analysis are documented in the online appendix.6
What Drives Cueing Effects Across Mainstream and Challenger Party Voters?
In thinking about the possible conditions that might lead different types of parties to influence opinion, we formulated two possible hypotheses (H3 and H4). The first one departs from the expressive interpretation of partisan leanings which emphasizes voters’ affinity with their own partisan groups. According to this, cueing appeals to affective and emotional motivations. The second, as we discussed, is more connected with the instrumental interpretation of partisanship which considers partisan leanings and its cueing effects as the product of party performance, especially as a result of its competence reputation.
We can observe looking at the overall results that both in-party group affinities and in-party reputation are two drivers of cueing when considering the whole set of voters relevant for this study, though with some differences (see Figure A5 in the online appendix). While in-party group affinities are relevant in boosting cueing effects only for one of the two issues (‘Reform of the Senate’), the moderating role of party reputation is consistent across both policies.7
However, this overall picture changes when we compare their moderating effects across voters of the different types of parties. Figure 3 displays how the role of in-party group affinities varies across voters for old mainstream and new challenger parties. We can clearly see that, for the ‘Senate’ issue, cueing effects tend to be stronger among subjects holding stronger affinities towards one’s own party group than those showing less favorable sentiments, especially when considering old mainstream party voters. Specifically, the cueing effect increases by 22% points in the case of party cues and by 19% points in the case of leader cues (in both p < 0.05; N = 480). Conversely, for the same issue, we do not find a moderating role of in-party group affinities when considering new challenger party voters. Finally, as found for the general voter sample, in-party group affinity is not relevant for cueing effects when looking at the issue of the ‘Health care and assistance for immigrants’, regardless of the type of party supporter (the results of the test of differences across average marginal effects can be found in Tables A13-A16 in the online appendix). Although this depends upon the issue at stake, cueing effects seems not be conditioned by the in-party group affinity, and this applies in particular to challenger parties. So, our third hypothesis (H3) seems to be unfounded.
Fig. 3
Party and leader cueing effects across old mainstream and new challenger parties and in-party group affinities. Note Lines across dots represents 95% confidence intervals. Estimates represents Average Marginal Effects of Logistic regression models predicting policy congruence given the assignment of partisan cues relative to the control (the dotted line)
Observing the results of Fig. 4, which helps us disentangle the moderating role of in-party reputation across the two groups of voters, we encounter a different scenario. For the issue ‘Reform of Senate’, party reputation exerts a very similar moderating role across groups of voters. Among mainstream voters cueing effects increase congruence between issue positioning and voting option by 22 (party labels) and 19 (leader labels) percentage points when respondents consider their own party more credible (p < 0.01, p < 0.05, respectively, N = 494), while among challenger voters cueing effects increase this congruence by 21 and 22% points under the same circumstances (p < 0.05, p < 0.01, respectively, N = 750). For the ‘Health care and assistance for immigrants’ issue, instead, the moderating role of our in-party issue credibility index seems to apply to challenger party voters only, but not to mainstream ones.8 For the former, cueing effects increase among respondents who believe their party to be more credible, for the latter the differences in cueing effects between respondents with varying assessments in terms of reputation are not statistically significant. Briefly, also our fourth hypothesis (H4) is not supported by data which point at the opposite direction, party issue credibility seems to have a stronger role in conditioning cueing among challenger party voters than among those opting for mainstream party options.
Fig. 4
Party and leader cueing effects across old mainstream and new challenger parties along in-party reputation. Note Lines across dots represents 95% confidence intervals. Estimates represents Average Marginal Effects of Logistic regression models predicting policy congruence given the assignment of partisan cues relative to the control (the dotted line)
To conclude, if we consider the two groups separately, we might say that while cueing among voters of mainstream parties is driven by the combination of both expressive and instrumental factors, for supporters of challenger parties cueing is conditional only to instrumental considerations, as specified in terms of party competence reputation.
Discussion
The emergence of challenger parties across many Western European democracies poses new questions as to the nature of cueing effects and the role of partisanship in shaping their effectiveness on preference formation (Brader et al., 2020; Pannico & Anduiza, 2022). In this article we revive the research on cueing effects across mainstream and challenger party voters, first by considering different sources of cueing. Indeed, leaders have become more important in the last few years and they might have been one of the relevant factors in the success of new, fringe and challenger parties. More importantly, we explore the conditional factors that may give strength to cues for different party voters. Our argument was that although new challenger parties may lack sufficient time to develop identification with groups and distinctive party reputations, they may still provide effective cues and reduce their competitive disadvantage developing affective social identity ties. We tested this employing an experimental design implemented in a democracy with the successful emergence of two challenger parties. This should, in principle, offer the perfect context for a robust analysis, helping the reliability of our conclusions. The findings have shown that partisan cueing, both from leaders and from parties themselves, is effective in improving voter policy congruence regardless of party type. Contrary to our expectations, while cueing effects among mainstream party voters seems to be a function of both expressive and instrumental partisanship, for challenger party voters cueing hinges on perceived party competence reputation.
We think that the preceding discussion and evidence offer several takeaways for scholars interested in partisanship, cueing effects and models of opinion formation. First, we show that new challenger parties might be as effective as mainstream parties in cueing individual preferences. This should imply that the role played by time and familiarity with parties in enhancing partisan cueing might have been overemphasized (Brader et al., 2012; Bullock, 2020). Second, we show that despite the importance given to leadership by the literature on personalization (Aarts et al., 2011; Costa Lobo & Curtice, 2015; Karvonen, 2010), leader cues are not more relevant than party cues even for new challenger parties. In this respect, one could argue that it might be hard to disentangle party signals from those provided by leaders, even in our design, since the two cues are inextricably linked. So, the differential importance of parties and leaders in shaping opinions deserves further attention in the future. With the data available, we were not able to test what happens when there are conflicting cues between parties and leaders.
Last, we show that the ability of new challenger parties and their leaders to change and influence public opinion is dependent on some instrumental assessments, despite the dominance affective reactions seem to have in politics nowadays. The fact that cueing effects apply to new parties and these are conditioned mainly by party evaluations merit more discussion. As for the Spanish case, the success of challenger parties may originate from “competence shocks” related to political scandals or periods of poor performance during economic downturn and crisis. In these scenarios, new challenger players might use anti-establishment rhetoric to contest mainstream party brand values (De Vries & Hobolt, 2020; Vidal, 2018). Thus, challenger parties build their reputation on their innovative character and attacking the status quo for its failure, such that their voters estimate party issue credibility or competence on a sort of “a-priori” evaluation rather than on actual government record or institutional experience. Thus, the role of evaluations in conditioning cueing effects among challenger parties somehow speaks also to the main approaches on how cues operate, the dual-process versus motivated reasoning approach (Bullock, 2020), contesting somehow the role of identities and affective attachments mostly advocated by the latter. At the same time, we cannot discard the possibility that other motivations lay beneath such competence evaluations, especially for challenger parties, potentially being the result of wishful thinking rather than the consideration of actual evidence. It would be an interesting question for future research whether the mechanisms that give strength to the effect of cues by challenger parties remains the same once those parties obtain a track-record in policymaking.
Acknowledgements
We would like to thank the anonymous reviewers for their helpful comments.
Declarations
Ethics Approval and Consent to Participate
All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and national research committees and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. This article does not contain any studies with animals performed by any of the authors.
Informed Consent
Informed consent was obtained from all individual participants included in the study.
Conflict of interest
The authors have no competing interest to declare that are relevant to the content of this article.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
