Consider your companions carefully: How voters perceive (coalition) governments’ poor valence images

Abstract

Numerous studies of comparative political behavior examine how voters perceive parties’ ideological positions on various policy issues, either in a standard uni-dimensional space or on single issues. These studies assume these ideological positions to be representative of the entire governing coalition, classifying the government as unitary. While a common assumption when assessing coalition governments’ ideological positions, it is unclear whether this logic of shared accountability holds for voters’ perceptions of valence. To fill this gap, I use a conjoint experiment to assess the perceptual influence of valence issues on coalitional accountability. Overall, my results show that unlike standard left-right ideological positions, voters project the prime minister’s and (junior) cabinet members’ low valence bidirectionally onto each other. This research has implications for the prime minister’s selection process for (junior) cabinet members and junior parties’ own calculus of whether to participate in a coalition or not.

Keywords

coalition politics voter perceptions valence issues conjoint experiment

Introduction

Recent scholarship demonstrates that the arrangement of coalition cabinet/governments often helps voters update their perceptions of cabinet members’ ideological positions and shifts (Fortunato and Adams, 2015; Fortunato and Stevenson, 2013). Given that a (coalition) cabinet is accountable to or, at a minimum, tolerated by the majority of a given parliament (Strøm, 2000), the formation of a (coalition) government generally contains politicians/parties who share similar ideologies so that policy promises can be fulfilled (Carroll and Cox, 2007; Martin and Stevenson, 2001). These relatively homogeneous cabinets, in turn, provide voters with an easily accessible political heuristic for determining the ideological positions of the cabinet as a whole and its member specifically.

As demonstrated in the literature, participating in a (coalition) cabinet is treated as a heuristic by voters when they try to avoid the prohibitive cost of monitoring actual policy-making and tend to follow less costly but informative and observable sources of information (Fortunato and Stevenson, 2013; Lau and Redlawsk, 2001). This is because rank-and-file voters generally recognize that government participants are both ideologically closer and more likely to compromise with each other (e.g., Dodd 1974; Martin and Stevenson, 2001; Schofield and Laver, 1985; Warwick, 1996). Therefore, cabinet members (and their parties) are considered to hold close/similar ideological positions by voters. In addition to the left-right positions, scholars also find that voters use the same heuristics to update their perceptions of cabinet members’ positions on non–left-right issues, such as the European Union (EU) integration policies, and responsibilities of policy outcomes (Adams et al., 2016; Duch et al., 2015; Spoon and Klüver, 2017).

Similar to the way voters use cabinet participation to update their perceptions of the ideological positions held by cabinet members, voters may also use the same cabinet participation heuristic to update their view of members’ valence evaluations. Since Stokes (1963), scholars have demonstrated that in addition to ideology, there is a second dimension, “valence,” that informs voters’ evaluations of politicians and parties to complement ideological explanations for political behavior and competition (Adams and Merrill, 2009; Stone and Simas, 2010).¹ Given that political scandals entail information of political-norm transgression (Thompson, 2013; Von Sikorski et al., 2020), there is a consensus among scholars that scandal involvement, in general, negatively affects politicians’ and parties’ electoral outcomes or political survival, as scandals diminish voters’ evaluation and trust in politicians’ and parties’ valence (e.g., Abney et al., 2013; Clark, 2009; Doherty et al., 2014; Funk, 1996). In other words, political scandals could be considered as a realized form of poor valence, given that scandals are inherently a breach of trust between voters and their elected elites. Thus, one may expect that sometimes a cabinet is dismissed due to transmission of one single case of a member’s poor valence—their involvement in a scandal—to other members.

Indeed, qualitative examples of contemporary politics demonstrate the possibility of this mechanism. For instance, consider the case of the former prime minister (PM) of Iceland, Sigmundur Davíe Gunnlaugsson. On 3 April 2016, the news media revealed about 11.5 million documents (the so-called “Panama Papers”) on how prominent leaders, politicians, and celebrities exploit the offshore tax regime, leaked from the database of the world’s fourth largest offshore law firm, Mossack Fonsecathe (Garside et al., 2016). From these documents, it was revealed that Sigmundur Gunnlaugsson and his partner, Anna Sigurlaug Pálsdóttir, were registered during November 2017 as the shareholders of a shell company, Wintris Inc (Obermaier and Obermayer, 2016). Later, in 2009, while Gunnlaugsson was the chairman of the Progressive Party having been elected to parliament that April, Gunnlaugsson failed to disclose these company shareholdings and instead sold these shares to his partner (Belfast Telegraph, 2016). This disclosure was required by a new transparency regulation which obliged Iceland’s members of parliament (MPs) to disclose any company shareholdings exceeding 25%. As a result of this perceived dishonesty and corruption, Gunnlaugsson stepped down as chairman on 5 April 2016 after widespread protests leading the interim government to call for early elections by fall within the same year (ABC News, 2016; Henley, 2016; Sigurdardottir, 2016).

Besides the Iceland scandal, other qualitative examples give some idea as to how a minister’s bad valence could endanger the entire survival of a cabinet. On 18 May 2019, Al-Serori et al., (2019) revealed a video to the public, which was secretly filmed before Austria’s election in October 2017. Based on that video, the Vice Chancellor of Austria and the leader of the Freedom Party at the time, Heinz-Christian Strache, was reported to have offered government contracts in return for campaign funds to a fake Russian broker, who had claimed that she was a niece of the Russian oligarch Igor Makarov (France 24, 2019; Reuters, 2019; Weise, 2019). On 18 May 2019, Heinz-Christian Strache announced his resignation and was later placed under criminal investigation due to this so-called “Ibizagate” or “Ibiza Affair” scandal (Heath and Karnitschnig, 2019; Huggler, 2019). This event not only hurt Mr. Strache’s own political career but also ended the current coalition government leading to a snap election in September 2019. This was due to a vote of no-confidence against Sebastian Kurz’s coalition cabinet which was passed and backed by both the Social Democrats and the Freedom party (Jones, 2019).

Since some PMs or ministers (and of course, their cabinet) do survive scandals in reality, these two events provide an ample reason to investigate how a PM’s or minister’s own poor valence, revealed through a scandal, can lead to the implosion of entire governments. In fact, they demonstrate that voters may project a cabinet member’s poor valence, regardless if that member is the PM or a minister, onto his/her colleagues, leading to a devaluation of the whole cabinet/government’s valence. Given this evidence, I investigate this theory of valence spillovers in cabinets by examining whether voters adopt cabinet participation to update their perceptions of cabinet members’ valence. Additionally, to operationalize valence, I focus specifically on political scandals, which are widely known to influence voters’ perceptions of character-based valence (e.g., Abney et al., 2013; Best et al., 2013; Clark, 2009; Clark and Leiter, 2014; Curini, 2015; Curini and Martelli, 2015).

In this paper, I argue that the negative effect of a cabinet member’s low valence can be transmitted from the PM to the minister and vice versa. I consider political scandals to be low-cost information given that first, political scandals merely carry information of moral transgression (Thompson, 2013); and second, voters tend to have identical views regarding politicians’ character-based images (Curini, 2015). In contrast to ideology or policy, voters are more likely to learn about both the PM’s or ministers’ scandals, and this in turn may create the room for “bidirectional” transmission of cabinet members’ poor valence within a cabinet. Nevertheless, the mechanisms of these transferals may be different. On the one hand, given that the PM usually dominates the attention of the media and is the most recognizable by voters (Duch et al., 2015; Fortunato and Adams, 2015; Klüver and Spoon, 2020), when the PM is embroiled in a scandal, rank-and-file voters update their perceptions of the whole cabinet’s valence—they lower their perceptions of other cabinet members’ valence based on their association with their embroiled colleague. On the other hand, as demonstrated by Strøm’s (2000)’s principal–agent theory, the PM is expected to be responsible for a given minister’s failure (including his/her failed valence image) due to the PM is the principal of the ministers while holding the status as the agent of the parliament who is responsible for the performance of the whole cabinet.

To examine whether voters causally project their perceptions of one specific cabinet member’s poor valence (i.e., his/her scandal involvement) onto other cabinet members, I employ a conjoint experiment, which presents all factors simultaneously to respondents and forces them to make evaluations based solely on the displayed factors. If a PM’s involvement in a scandal significantly and negatively impacts voters’ choices of a minister (and vice versa), this result would demonstrate that voters do transfer their perceptions of one cabinet member’s low valence to other cabinet individuals—that is, they expect cabinet members to share similar levels of low valence. Figure 1 below explains the flow of causal effects I examine in this experiment.

Figure 1.

Estimated effects from the two experiments.

The conjoint experiment results provide strong evidence for my argument that voters do utilize cabinet participation as a type of heuristic to update their evaluations of cabinet members. I find that unlike non-valence (ideological) perspectives (See Duch et al., 2015; Fortunato and Adams, 2015; Fortunato and Stevenson, 2013), voters project a junior member’s poor valence onto the PM and their party, just as they project the PM’s bad valence onto the junior party. Furthermore, these causal effects are homogeneous across unified or coalition governments. The findings complement our existing understanding surrounding how voters utilize heuristics to update their perceptions of politicians’, parties’, or governments’ various characteristics, including valence. Additionally, understanding this mechanism can help us better explain why and how a government’s survival could be terminated sometimes and thus better predict when a snap election will happen, given an outbreak of a cabinet member’s scandal involvement.

(Coalition) Governments and their perceived valence

Research on public opinion has consistently demonstrated that rank-and-file voters are relatively uninformed about politics due to the high cost, and low benefit, associated with acquiring complex information about political actors’ ideology and policy stances (e.g., Bartels 1996; Converse, 2006; Zaller, 1992). In fact, scholars find that voters, using party manifestos, only weakly, if at all, update their perceptions of parties’ policy positions after they have, in fact, shifted (Adams and Somer-Topcu, 2009; Adams et al., 2011a). However, these findings do not necessarily imply that general voters are completely ignorant about politics; rather, they tend to rely on low- or no-cost heuristics, such as party affiliation, to gather information about political actors (Bernhard and Freeder, 2020; Dancey and Sheagley, 2013).

Additionally, in the context of European politics and parliamentarism, recent studies show that European rank-and-file voters use another type of heuristic to inform themselves: a politician or party’s participation in a (coalition) cabinet (Fortunato and Stevenson, 2013). As the literature has demonstrated, in general, politicians and parties are more likely to form a (coalition) government together when their positions are closer ideologically (e.g., Golder, 2006; Martin and Stevenson, 2001; Schofield and Laver, 1985). Furthermore, politicians and parties within a (coalition) government are also more likely to compromise (i.e., merge) their positions (e.g., Dodd, 1974; Martin and Vanberg, 2005; Warwick, 1996). Thus, joining a (coalition) cabinet conveys relatively accurate ideological information and has a low cost for observation. To this point, empirical studies consistently find evidence of voters using cabinet heuristics to update their perceptions of politicians’/parties’ shifts of ideological positions, shifts of policy stances, and policy accountability (Adams et al., 2016; Duch et al., 2015; Fortunato and Adams, 2015; Fortunato and Stevenson, 2013; Spoon and Klüver, 2017).

Since Stokes’s (1963) foundational work on ideological spaces, research has consistently found that in addition to left-right ideology and policy positions, voters also rely on another important source of information to make decisions: the valence of a politician or party. Generally, valence issues are divided into two categories by scholars: “issue ownership” (policy-based) and “character-based” (non–policy-based) (Curini, 2015). The former refers to how competent a politician is in “handling” a certain issue or how much expertise they have regarding a given issue (Clarke et al., 2004), for example, liberal politicians are considered to be better on welfare issues; the latter refers to politicians’ character-based images, often informed by perceptions of trustworthiness, integrity, etc. (Clark, 2009). For the purposes of this article, I focus on character-based valence, including but not limited to honesty, competence, charisma, likability, trust, and unity (Adams, 2012). These characteristics have consistently been found to alter voters’ evaluations of politicians and parties and influence their vote-choice (e.g., Adams and Merrill, 2009; Clark, 2009; Clark and Leiter, 2014; Groseclose, 2001). Simply put, because valence encompasses the non-policy qualities that citizens value in their elected representatives and parties, citizens ultimately desire to have high-valence representatives whom they can trust not to shirk their political responsibilities (Mondak, 1995; Stone and Simas, 2010). In other words, voters tend to hold politicians/parties to be accountable for their own valence images. Therefore, given that rank-and-file voters tend to utilize cabinet participation as an informational heuristic to update their perceptions of the participating members’ ideological positions, I argue that they will also utilize this same heuristic to update their perceptions of (coalition) cabinet members’ valence.

To operationalize the concept of poor valence, I concentrate on political scandals, a widely utilized variable that indicates a reduction in a politician or party’s perceived valence (e.g., Abney et al., 2013; Best et al., 2013; Clark, 2009; Clark and Leiter, 2014; Curini, 2015; Curini and Martelli, 2015). In fact, political scandals tend to expose behavior that implies dishonesty, corruption, and/or incompetence, which should significantly decrease valence-based evaluations of the embroiled politician. For instance, politicians who misreport funds are often considered to be corrupt and dishonest. This type of behavior can signal representational shirking to citizens, violating the public trust between elected representatives and their citizens; other scandalous issues, such as sexual harassment, are often considered as a violation of integrity and the moral values held by most citizens.

Furthermore, because political scandals are usually presented in the media as straightforward narratives describing politicians’ moral transgression (Thompson, 2013), on which voters usually have identical views (Curini, 2015), I expect that political scandals should be considered as a low-cost form of low valence since they are nothing more than negative stories to the public. Given the combination of being simple and providing important valence information, I expect scandals to be a significant type of informational heuristic through which voters evaluate politicians and parties, which is supported by previous literature (e.g., Ares and Hernández, 2017; Doherty et al., 2014; Dewan and Dowding, 2005; Dewan and Myatt, 2007; Ecker et al., 2016; Kumlin and Esaiasson, 2012).

Given this context, I argue that when any member of a (coalition) cabinet is newly perceived as having low valence (i.e., is involved in a scandal), rank-and-file voters will perceive other members as having low valence as well. Unlike policy positions that voters project unidirectionally from the PM onto other (junior) cabinet members (Fortunato and Adams, 2015; Klüver and Spoon, 2020), or inversely, that voters only project (junior) ministers’ accountability on to the PM (Duch et al., 2015), I hypothesize that voters not only negatively update their perceptions of ministers (in a single-party government) and junior members (in a coalition government) in line with the PM’s scandal involvement, but also negatively update their perceptions of the PM’s valence when a minister or junior member is embroiled as well.

Note that these are two types of “collective responsibility” (i.e., ideology and valence) in that they differ in voters’ information updating mechanism. Given that ideology and policy (accountability) are “difficult” information (Converse, 2006), rank-and-file voters cannot efficiently trace and evaluate every politician and party’s ideology or policy positions (Adams et al., 2011a; Fernandez-Vazquez, 2014). Therefore, they tend to first (correctly or incorrectly) recognize the ideology or policy positions of the PM or the senior coalition party, who dominate both the media’s attention and government-recognition, and then further calibrate junior members’ position, but not vice versa (Duch et al., 2015; Fortunato and Adams, 2015; Fortunato and Stevenson, 2013; Klüver and Spoon, 2020). On the contrary, given that it only carries low-cost information, voters are much more likely to learn about a politician’s scandal, regardless of whether it is the PM or a minister, and this could lead voters to perceive that other cabinet members have low valence too. In other words, this cabinet-level valence heuristic works bidirectionally: the PMs’ acquisitions of valence images influence those of other ministers and junior members, and vice versa.

On the one hand, given the PM’s status as the head of their party and government, voters should be more likely to consider the PM’s low valence as representative of the whole cabinet’s valence and thus negatively affect the evaluations of lower cabinet (junior) members’ valence. While (junior) ministers can always choose to exit the cabinet when they disapprove of the incident (Fortunato and Adams, 2015), staying in the cabinet while the PM is involved in a scandal could be considered as an endorsement of the event by voters. On the other hand, based on Strøm’s (2000) argument regarding principal–agent relationships in parliamentary systems, because the PM is the agent of the parliament under the “chain of delegation,” any cabinet member’s scandal involvement signals that PM’s inability to manage agency loss within the government. In other words, in addition to their own valence performance, the PM is expected to hold accountability for their agents, that is, the rest members of the cabinet, too.

I ignore the interdependent valence analysis between cabinet ministers (not involving the PM) in this project. Based on the principal–agent model, the structure of the chain of delegation implies that the relationships between ministers are indirect/secondary—a minster is linked to the other ministers within a cabinet only through their common linkages toward the PM. One can analogize such relationships as siblings under a family tree. In that vein, the results of the interdependent valence assessments between cabinet ministers only demonstrate how the negative effect of one minister’s scandal is depleted while it is transmitted along with this indirect route but not the direct impact of a minister’s scandal onto voters’ evaluations of other members. Overall, I argue that once any cabinet member (including the PM) is involved in a scandal, the resulting poor valence image not only hurts the offending cabinet member’s evaluations, but also hurts voters’ perception of other cabinet members. Based on these theories, I will test the two following hypotheses:

When the PM is involved in a political scandal, this will negatively affect voters’ preferences toward their ministers, under both uniform and coalition cabinets.

When a minister is involved in a political scandal, this will negatively affect voters’ preferences toward their PM, under both uniform and coalition cabinets.

Design and data

To examine the hypotheses derived from the theoretical framework, I adopt a conjoint experimental design, which provides two key advantages. First, internal validity is preserved through randomization. This is important because voters’ evaluation regarding politicians’ scandalous events are frequently influenced by external factors, such as voters’ preexisting preferences or how media reports the events, which further impede researchers’ ability to examine voters’ “immediate reaction” to scandals (i.e., the “pure effects” of scandals will be biased) (Doherty et al., 2011; Funk, 1996). For instance, citizens may disregard (stress) the negative scandal of a politician who is (not) from the same party as such information contradicts (corresponds) to their predisposition (Bolsen et al., 2014; Zaller, 1992); or for another instance, the way media frames and reports scandals, either positively or negatively, could alleviate or deepen the perceived negative magnitude of scandals (Galvis et al., 2016; Puglisi and Snyder, 2011).

In this design, random assignment is achieved by randomly assigning a number of different treatments that are presented in a list form. Furthermore, at least two lists are assigned, which allows participants to pick the list they prefer or to rank all the lists. In this project, since each participant randomly received two lists combining politicians’ attributes across all possible levels of comparison, the estimated causal effects of each treatment should be unbiased, including the measurement of the effect of a certain cabinet member’s involvement of bad valence or political scandal on voters’ preferences toward other members.

Second, in contrast to survey experiments, the conjoint analysis experiment can more accurately resemble the real-world scenario in which voters usually consider multiple aspects of information simultaneously to make decisions. Under the traditional experimental design, participants are asked to respond to one single manipulation, and researchers then estimate the treatment’s causal effects through the results. However, this design is limited because researchers are unable to compare different treatments (Hainmueller and Hopkins, 2015; Hainmueller et al., 2013). In a conjoint design, respondents are forced to evaluate multiple treatments at the same time. This allows researchers to carefully examine how respondents weigh different treatments. In other words, a conjoint experiment can help us further determine whether these causal effects of valence are trivial or not, as compared to other factors (i.e., partisanship and policy stances).

The study’s data were collected in 2018 with the help of the EssexLab at the University of Essex. All participants were British citizens and randomly selected from a diverse sample pool recruited by the EssexLab in advance.² In addition, 120 of 241 random participants (from the same recruited sample pool) who did not take the survey in the lab on any session were instead invited to take the survey remotely.³ In general, the in-lab respondents were younger, relatively liberal (in terms of partisanship and attitudes toward the EU), and had higher levels of education. Including the online sample in the analysis helps me more or less increase the samples’ diversity (especially respondents who are not college students).

In this experiment, respondents were asked to make a choice between a pair of PMs or cabinet ministers several times. After a short introduction explaining the experiment’s purpose, each respondent faced two rounds of decision-making tasks and one round of demographic questions. At the beginning of each round of decision-making tasks, respondents were provided with a sample profile, as presented in Figure 2, in which 2a represents the round selecting PMs and 2b the round selecting ministers.⁴ This design allows me to measure how specific types of PMs or ministers are evaluated by voters through analyzing the chance that they are selected in each decision-making task.

Figure 2.

Instructions for interpreting profiles.

In the decision-making rounds, respondents made decisions about two different types of politicians separately, namely, the PM and minister. As shown in the instructions in Figure 2, respondents were asked during the experiment to select their preferred PMs in the first round and their preferred ministers in the second round. Similar to Hainmueller et al., (2013), in each round, respondents were presented with six decision tasks. More precisely, during every evaluation, each respondent will be presented with all attributes—including gender, partisanship, policy position, and scandal involvement—of a (unspecified) cabinet member or the PM, so that the respondent can make an evaluation/choice regarding the current pair of ministers or PMs. To minimize the potential bias created by the order in which the attributes were shown, the attributes were reordered randomly every two tasks. Following the instructions, respondents were presented with profiles of two PMs or two ministers on their screens in each round and were asked to select their preferred profile each time. Figure 3 features the samples questions for each round. This design could increase the sample size of the experiment through treating each unique response as a “different respondent” given that respondents receive different treatments with different combinations and orders of features every time.

Figure 3.

Sample questions.

Note that in this experiment, I do not treat the concept of political scandal as one general concept or merely focus on one specific aspect of scandals. Rather, respondents are allowed to consider different aspects of scandals to assess heterogeneity in the effect of various types of scandals. Indeed, previous literature has distinguished political scandals into two general categories—private scandals (also known as moral or sexual scandals, which are related to sexual misconduct, especially extramarital affairs) and public scandals (also known as financial scandals, which include misuse of allowance, tax evasion, and embezzlement), and demonstrated their differing levels of impacts (e.g., Carlson et al., 2000; Doherty et al., 2011, 2014; Funk, 1996; Sarmiento-Mirwaldt et al., 2014; Von Sikorski, 2018). Considering the balance between validity and dimensionality, in addition to the two traditional types, I also include sexual harassment, which is largely ignored by the standing literature and has become more salient among the public recently (Costa et al., 2020), and plagiarism, which has commonly happened to high-profile European politicians (e.g., the former German defense minister, Karl-Theodor zu Guttenberg (see Gathmann (2011) for detailed coverage), the former Hungarian president, Pal Schmitt (see Karasz (2012) for detailed report), etc.). By including multiple types of scandals, this design not only simulates the experimental environment of decision-making with the reality voters face, but also further increases the validity of this experiment. Indeed, the inclusion of multiple scandal types provides for a relatively comprehensive analysis regarding the impacts of varying scandals.

Table 1 presents the attributes that were included in this experiment. To make sure that the treatments were sufficiently realistic to participants, two specific constraints were imposed in each round: the PM can only be from either the Conservative Party or Labour Party and the minister is only allowed to be a Liberal Democrat when the PM is a Conservative (so as to preserve the realism of a potential coalition government). Overall, there were 1347 profiles in the first round and 4047 profiles in the second round. Each profile had an equal chance of being assigned to the participant. Based on the pre-test experience, each participant was provided with a handout featuring an (online) index (see Supplemental Appendix A) of several potentially confusing terms—such as “plagiarism”—in case they had a hard time imagining those specific scandals.

Table 1.

The list of featured attributes and levels by rounds.

	Attributes	Levels
Prime minister	Gender	Female
	Gender	Male
	Party affiliation	Conservative Party
	Party affiliation	Labour Party
	Position on EU integration	Support EU integration
		Neutral
		Oppose EU integration
	Scandal	No scandal
		Sexual harassment
		Misreporting funds
		Plagiarism
		Having an extramarital affair
	Party affiliation (minister)	Conservative Party
		Labour Party
		Liberal Democrats
	Scandal (minister)	No scandal
		Sexual harassment
		Misreporting funds
		Plagiarism
		Having an extramarital affair
Minister	Gender	Female
	Gender	Male
	Party affiliation	Conservative Party
		Labour Party
		Liberal Democrats
	Position on EU integration	Support EU integration
		Neutral
		Oppose EU integration
	Scandal	No scandal
		Sexual harassment
		Misreporting funds
		Plagiarism
		Having an extramarital affair
	Party affiliation (PM)	Conservative Party
	Party affiliation (PM)	Labour Party
	Position on EU integration (PM)	Support EU integration
		Neutral
		Oppose EU integration
	Scandal (PM)	No scandal
		Sexual harassment
		Misreporting funds
		Plagiarism
		Having an extramarital affair

Finally, the analytical strategy adopted in this paper is not to estimate (conditional) average marginal component-specific effects (AMCEs) as used in most recent political science conjoint analyses (e.g., Campbell et al., 2019; Clayton et al., 2019; Gutiérrez-Romero and LeBas, 2020; Teele et al., 2018; Wright et al., 2016). Instead, I calculate marginal means (MMs), which is the percentage of times respondents choose profiles given a specific category across all possible combinations of categories of the rest of features/variables (Hobolt et al., 2020). I use this approach for two reasons: first, to avoid the potential issue of distorted interpretation given that AMCEs cannot be directly compared, and this is because AMCEs “do not provide information about absolute levels of favorability toward profiles with each feature” (Clayton et al., 2019; Leeper et al., 2020); second, to avoid the potential complexity of calculating weighted AMCEs based on the feature restrictions/constraints which eliminate the randomization of treatments (Hainmueller et al., 2013).

Analysis

I present the analytical results for the first hypothesis in Figure 4 below. First of all, each point and the corresponding error bar in Figure 4(a) is the mean probability and the 95% confidence bar that a feature’s causal impact on a voter’s choice of a certain minister.⁵ As five estimates of Figure 4(a) demonstrate, in general, I find that compared to the minister whose PM is not involved in a sexual harassment or misreporting funds scandal, voters significantly downgrade the minister whose PM is involved in those types of scandals. More precisely, when the PM is involved in sexual harassment, their minister is around 30% less likely to be picked by respondents, compared with one who has no scandals at all. Furthermore, when the PM is involved in any other scandal (i.e., misreporting funds, plagiarism, and having extramarital affair), his minister is between 10% and 20% less likely to be picked by the respondents. In other words, voters’ perceptions of a minister’s valence will be negatively affected by such incidence.

Figure 4.

Estimates of respondents’ evaluations of ministers.

According to Supplemental Figure C1 in Supplemental Appendix C, the estimated effects of other features further validate the results. First, as described earlier, the recruited respondents lean more toward the Labour Party and are ideologically more liberal, and the effects of the minister’s own features reflect this structure as well—the respondents significantly prefer Labour over Conservative ministers (about 5% more likely) and significantly dislike ministers who are Eurosceptics, that is, those who oppose EU integration (about 10% more likely). In addition, the results demonstrate that voters significantly devalue ministers with low valence images (i.e., those involved in scandals), which is consistent with findings from previous studies (e.g., Abney et al., 2013; Clark 2009; Curini 2015; Dewan and Dowding 2005). Furthermore, Supplemental Figure C1 demonstrates that how voters evaluate a minister is causally affected by the policy position of that minister’s PM: voters tend to perceive ministers’ policy positions to be similar with that of their PM so they utilize the PM’s position regarding EU integration as a heuristic to evaluate them. This result comports directly with the findings of Adams et al. (2016).

Ultimately, this conjoint design helps reduce priming effects, such as social desirability bias (SDB), through multiple treatment assignments (Campbell et al., 2019; Gutiérrez-Romero and LeBas, 2020; Hankinson, 2018; Teele et al., 2018). Indeed, the simultaneous presentation of information should effectively distract respondents from focusing on one specific factor (Gutiérrez-Romero and LeBas,2020; Teele et al., 2018). As such, I expect the conjoint analysis experiment to mitigate the potential SDB of political scandals by simultaneously randomizing multiple factors presented to respondents in a given experimental frame. In addition, given that conjoint analysis estimates the relative importance among factors (Bansak et al., 2016), magnitude of scandals are indeed consistent with conventional wisdom that voters tend to pay greater attention to relatively more accessible information.

I further examine the second half of the first hypothesis as to whether the negative impacts of a PM’s bad valence and scandal involvement on voters’ evaluation of cabinet ministers are significantly different under single-party or coalition cabinets in Figure 4(b). To do so, I keep the preference tasks only for the PMs who are from the Conservative Party given the constraint of the design and the reality that the Labour Party has never formed any coalition cabinets. In Figure 4(b), the red points and bars represent the estimates of the single-party (Conservative) cabinets and the blue points and bars represent the estimates of the coalition cabinets in which the Liberal Democratic Party is junior to the Conservatives. Given that the effects of the PM’s poor valence and scandals are not significantly different (i.e., the 95% bar of one color contains the point of the other color), I conclude that the negative causal impact of the PM’s acquisition of low valence image (i.e., scandal involvement) on cabinet members’ valence images is independent of the type of government (single-party vs coalition). In other words, voters attribute lower valence to cabinet ministers regardless of if they are from the same or junior party. As a validity check, the findings of Figure 4(b) also show that the impacts of a minister’s own personal attributes are statistically the same over different types of cabinets.

I next present the analytical results for the second hypothesis in Figure 5. As demonstrated by the five estimates in Figure 5(a), the first half of the second hypothesis is confirmed in that voters’ evaluations/perceptions of PMs are significantly and negatively affected by acquisitions of poor valence images (i.e., scandal involvement) of ministers from the same cabinet. Similar to the results in Figure 4(a), I find that when a minister is involved in sexual harassment, his PM is around 35% less likely to be picked by respondents, compared with one who has no scandal involvement; while a minister is involved in any of other scandals, their PM is between 10% to 20% less likely to be picked by the respondents.

Figure 5.

Estimates of respondents’ evaluations of the PMs.

As a validity check for the results of this round, the estimated effects of the PM’s own personal attributes are consistent with the structure of the sampled respondents and my expectation. For ideology and policy related features, Supplemental Figure C3 in Supplemental Appendix C demonstrates that respondents prefer Labour PMs (about 10% more likely) and dislike ones who oppose EU integration (about 15% more likely). These results are highly consistent with those that I found previously, summarized in Supplemental Figure C1. Furthermore, I also find the expected and simple relationship that a PM’s own scandal negatively affects how voters evaluate them. This indicates that voters, in general, do not prefer politicians who are involved in scandals or holding poor valence images, which is consistent with valence theory and previous findings.

Finally, I examine the potential different effects of acquired lower valence by a junior party cabinet member versus a single-party cabinet on the updated valence image of their PMs. As presented in Figure 5(b), I confirm the second half of the second hypothesis that the causal effect of a minister’s acquisition of poor valence is statistically the same under the single-party or coalition governments. Indeed, as the figure shows, the 95% bar of the red color (the single-party, Conservative cabinet) contains the point of the blue color (the coalition cabinet, Conservative and Liberal Democrats), and vice versa. Overall, unlike ideology or policy accountability, which only influences voters’ perceptions unidirectionally—the PM represents cabinet ministers but not vice versa (Duch et al., 2015; Fortunato and Adams, 2015), I find a bidirectional relationship: voters do in fact consider both the PM’s and (junior) minister’s acquisition of poor valence images as indicative of the cabinet ministers’ and PM’s poor valence, respectively.

Conclusion and discussion

Previous studies of parliamentary politics have consistently found that instead of using complex information such as party manifestos, roll-call votes, and parliamentary debates, voters rely more on simple, low-cost observable actions—cabinet participation—to inform themselves of politicians’ and parties’ left-right ideology, policy positions and shifts, and policy accountability (Duch et al., 2015; Fortunato and Adams, 2015; Fortunato and Stevenson, 2013). In this article, I extend the literature to examine whether voters utilize the same heuristics to update their perceptions of politicians’ and parties’ valence. I find, in line with previous ideological studies, that voters use participation in (coalition) cabinets to determine cabinet members’ valence images.

More precisely, when the PM is involved in a scandal, I find that voters project this PM’s poor valence image onto their ministers. In addition, unlike ideology and policy stances, I find a bidirectional relationship: when a minister is involved in a scandal, voters project this minister’s poor valence image onto his PM as well. Moreover, I find that the impact of scandal involvement is statistically equal under different types of cabinets. In other words, the magnitude of the impact of the PM or (junior) minister’s poor valence image onto their (junior) minister or PM, respectively, does not change between single-party and coalition governments. By and large, the findings suggest that such poor valence images will be transmitted vertically within a cabinet (i.e., from the PM to ministers, and vice versa), and furthermore, they will also be transmitted horizontally within a cabinet (i.e., from the senior party to the junior one, and vice versa). These results add to our knowledge as to why sometimes a minister’s personal scandal can endanger the survival of the entire cabinet.

Importantly, these results suggest several limitations and directions for future research. First, the derived results are limited to the specific setting of the U.K.’s cabinet, given that: a) this experiment only considers three major parties and a simple situation under which a coalition government only includes two parties at most each time; b) the assumption of principal–agent relationships in this experiment are not necessarily that clear in other parliamentary democracies, such as the Italian or Israeli cabinet; c) the extent to which British voters’ ability of distinguishing coalition and majority governments remains somewhat unclear given that they have only experienced a coalition government once (from 2010 through 2015) since WWII. All these, together, highly differentiate the U.K. case with most coalition governments in Europe, and researchers need to be cautious not to over-generalize the results to other countries.

Second, this experiment was conducted amidst a significant increase in awareness across advanced democracies surrounding the sexual harassment and assault of women by men in positions of power, including those in politics. This awareness increased following the rise of the 2017 “#MeToo” movement generally, and specifically following significant events across developed democracies (e.g., Costa et al., 2020; De Benedictis et al., 2019; Verge, 2020). Given that the salience of sexual harassment and assault peaked at an unprecedented level in the U.K. at the time of the experiment, the impact of the #MeToo movement may have either temporarily or permanently increased effect of the sexual harassment scandal on valence evaluations, compared with previous time-periods or other regions (Cameron et al., 2009; Fisman and Miguel, 2007).

Third, given that voters tend to view politicians who actively, and positively, interact with each other as being politically closer (Adams et al., 2021), the degree of the spillover effects between them might be amplified. As we have often witnessed, in most coalition governments, party elites in a governing coalition often refrain from publicly criticizing each other, both in terms of policy disagreements but also on valence grounds. Therefore, if coalition partners publicly defend each other in cases where one of them is under attack for poor valence, these perceived closer relationships may prompt citizens to escalate the magnitude of spillover effects to a greater extent.

Fourth, the potential conditional effects are not examined due to the scope of this paper and space limitations. Previous literature shows that voters’ evaluation of political scandals are conditioned by politicians’ or voters’ personal attributes, such as gender, party affiliation, or even ideology (e.g., Barnes and Beaulieu, 2014; Graham et al., 2009; Saxton and Barnes, 2020). In that vein, one may expect that the spillover effects of scandals among cabinet members are conditional as well.

Due to the fact that experimental methods are naturally limited in their ability to be generalized, future studies should replicate or build upon this project under different settings so as to better understand the mechanisms involved in valence evaluations related to political scandals. Nevertheless, because the issue of scandals within a cabinet is usually referred to as a typical moral hazard problem where agents shirk their responsibility under parliamentary democracies (Dewan and Myatt, 2007), the empirical analysis provides conservative insights regarding how the negative impact of scandals could be transmitted under a multiparty system and serves as a step stone for analyses of more complex cases.

Finally, my overall findings have two major implications. First, when heading a single-party government, the PM not only needs to carefully screen cabinet members who will not fail on their policy promises, but also who will not fail to uphold positive valence images; while under a coalition government, screening out junior members who will fail to uphold positive valence images should be largely considered by the PM and the senior party given that junior members’ poor valence images will affect voters’ evaluations of the whole cabinet.

Furthermore, the junior party should participate in a coalition only when the senior party’s position is as close to its own as possible because of future electoral disadvantages caused by voters (mis)perceiving the junior party’s position as in line with the senior party’s (Fortunato and Adams, 2015; Klüver and Spoon, 2020). More importantly, given that parties tend to change their positions while having poor valence images (Adams et al., 2011b; Clark, 2014; Groseclose, 2001), when the PM and senior party are perceived to have poor valence images, they may change and push away their positions from the junior parties, which will exacerbate these disadvantages. In combination with previous findings, these results suggest that junior parties need to carefully consider both the positions and valence images of the PM and senior party when deciding whether to join a coalition or not.

Supplemental Material

sj-pdf-1-ppq-10.1177_13540688211043472 – Supplemental Material for Consider your companions carefully: How voters perceive (coalition) governments’ poor valence images?

Supplemental Material, sj-pdf-1-ppq-10.1177_13540688211043472 for Consider your companions carefully: How voters perceive (coalition) governments’ poor valence images? by Tzu-Ping Liu in Party Politics

Footnotes

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.

ORCID iD

Tzu-Ping Liu

Supplemental material

Supplementary material for this article is available online.

Notes

Author biography

Tzu-Ping Liu is an assistant professor at the University of Taipei. His research focuses on European politics, political behavior, and political methodology.

References

ABC News (2016) Panama Papers: Iceland PM Gunnlaugsson Urged to Resign Amid Mossack Fonseca Data Leak. New York, NY: ABC News. Available at: https://www.abc.net.au/news/2016-04-04/iceland-pm-urged-to-resign-amid-panama-papers-scandal/7295742.

Abney

Adams

Clark

, et al. (2013) When does valence matter? Heightened valence effects for governing parties during election campaigns. Party Politics 19(1): 61–82.

Adams

(2012) Causes and electoral consequences of party policy shifts in multiparty elections: theoretical results and empirical evidence. Annual Review of Political Science 15: 401–419.

Adams

Somer-Topcu

(2009) Moderate now, win votes later: the electoral consequences of parties’ policy shifts in 25 postwar democracies. The Journal of Politics 71(02): 678–692.

Adams

Merrill

III (2009) Policy-seeking parties in a parliamentary democracy with proportional representation: a valence-uncertainty model. British Journal of Political Science 39(03): 539–558.

Adams

Ezrow

Somer-Topcu

(2011a) Is anybody listening? Evidence that voters do not respond to European parties’ policy statements during elections. American Journal of Political Science 55(2): 370–382.

Adams

Merrill

III Simas

, et al. (2011b) When candidates value good character: a spatial model with applications to congressional elections. The Journal of Politics 73(1): 17–30.

Adams

Ezrow

Wlezien

(2016) The company you keep: how voters infer party positions on European integration from governing coalition arrangements. American Journal of Political Science 60(4): 811–823.

Adams

Weschle

Wlezien

(2021) Elite interactions and voters’ perceptions of parties’ policy positions. American Journal of Political Science 65(1): 101–114.

10.

Al-Serori

Das Gupta

Munch

, et al. (2019) Caught in the trap. Available at https://projekte.sueddeutsche.de/artikel/politik/caught-in-the-trap-e675751/.

11.

Ares

Hernández

(2017) The corrosive effect of corruption on trust in politicians: evidence from a natural experiment. Research & Politics 4(2): 1–8.

12.

Bansak

Hainmueller

Hangartner

(2016) How economic, humanitarian, and religious concerns shape European attitudes toward asylum seekers. Science 354(6309): 217–222.

13.

Barnes

Beaulieu

(2014) Gender stereotypes and corruption: how candidates affect perceptions of election fraud. Politics & Gender 10(3): 365–391.

14.

Bartels

(1996) Uninformed votes: information effects in presidential elections. American Journal of Political Science 40(1): 194–230.

15.

Belfast Telegraph (2016) Iceland’s pm says he will not resign in panama papers scandal. Belfast Telegraph. Available at: https://www.belfasttelegraph.co.uk/news/panama-papers/icelands-pm-says-he-will-not-resign-in-panama-papers-scandal-34597940.html.

16.

Bernhard

Freeder

(2020) The more you know: voter heuristics and the information search. Political Behavior 42(2): 603–623.

17.

Best

Ladewig

Wong

(2013) Owning valence issues: the impact of a “culture of corruption” on the 2006 midterm elections. Congress & the Presidency 40(2): 129–151.

18.

Bolsen

Druckman

Cook

(2014) The influence of partisan motivated reasoning on public opinion. Political Behavior 36(2): 235–262.

19.

Cameron

Chaudhuri

Erkal

, et al. (2009) Propensities to engage in and punish corrupt behavior: experimental evidence from Australia, India, Indonesia and Singapore. Journal of Public Economics 93(7–8): 843–851.

20.

Campbell

Cowley

Vivyan

, et al. (2019) Legislator dissent as a valence signal. British Journal of Political Science 49(1): 105–128.

21.

Carlson

Ganiel

Hyde

(2000) Scandal and political candidate image. Southeastern Political Review 28(4): 747–757.

22.

Carroll

Cox

(2007) The logic of Gamson’s law: pre-election coalitions and portfolio allocations. American Journal of Political Science 51(2): 300–313.

23.

Cartwright

Ramalingam

(2019) Framing effects in public good games: Choices or externalities? Economics Letters 179: 42–45.

24.

Clark

(2009) Valence and electoral outcomes in Western Europe, 1976-1998. Electoral Studies 28(1): 111–122.

25.

Clark

(2014) Understanding parties’ policy shifts in Western Europe: the role of valence, 1976–2003. British Journal of Political Science 44(02): 261–286.

26.

Clark

Leiter

(2014) Does the ideological dispersion of parties mediate the electoral impact of valence? A cross-national study of party support in nine Western European democracies. Comparative Political Studies 47(2): 171–202.

27.

Clarke

Sanders

Paul

, et al. (2004) Political Choice in Britain. Oxford, UK: Oxford University Press on Demand.

28.

Clayton

Ferwerda

Horiuchi

(2019) Exposure to immigration and admission preferences: evidence from France. Political Behavior 47(1): 1–26.

29.

Converse

(2006) The nature of belief systems in mass publics (1964). Critical Review 18(1–3): 1–74.

30.

Costa

Briggs

Chahal

, et al. (2020) How partisanship and sexism influence voters’ reactions to political# metoo scandals. Research & Politics 7(3): 2053168020941727.

31.

Curini

(2015) The conditional ideological inducement to campaign on character valence issues in multiparty systems. Comparative Political Studies 48(2): 168–192.

32.

Curini

Martelli

(2015) A case of valence competition in elections. Party Politics 21(5): 686–698.

33.

Dancey

Sheagley

(2013) Heuristics behaving badly: party cues and voter knowledge. American Journal of Political Science 57(2): 312–325.

34.

De Benedictis

Orgad

Rottenberg

(2019) # Metoo, popular feminism and the news: a content analysis of UK newspaper coverage. European Journal of Cultural Studies 22(5–6): 718–738.

35.

Dewan

Dowding

(2005) The corrective effect of ministerial resignations on government popularity. American Journal of Political Science 49(1): 46–56.

36.

Dewan

Myatt

(2007) Scandal, protection, and recovery in the cabinet. American Political Science Review 101(1): 63–77.

37.

Dodd

(1974) Party coalitions in multiparty parliaments: a game-theoretic analysis. American Political Science Review 68(3): 1093–1117.

38.

Doherty

Dowling

Miller

(2011) Are financial or moral scandals worse? It depends. PS: Political Science & Politics 44(4): 749–757.

39.

Doherty

Dowling

Miller

(2014) Does time heal all wounds? Sex scandals, tax evasion, and the passage of time. PS: Political Science & Politics 47(2): 357–366.

40.

Duch

Przepiorka

Stevenson

(2015) Responsibility attribution for collective decision makers. American Journal of Political Science 59(2): 372–389.

41.

Ecker

Glinitzer

Meyer

(2016) Corruption performance voting and the electoral context. European Political Science Review 8(3): 333–354.

42.

Fernandez-Vazquez

(2014) And yet it moves. Comparative Political Studies 47(14): 1919–1944.

43.

Fisman

Miguel

(2007) Corruption, norms, and legal enforcement: evidence from diplomatic parking tickets. Journal of Political Economy 115(6): 1020–1048.

44.

Fortunato

Adams

(2015) How voters’ perceptions of junior coalition partners depend on the prime minister’s position. European Journal of Political Research 54(3): 601–621.

45.

Fortunato

Stevenson

(2013) Perceptions of partisan ideologies: the effect of coalition participation. American Journal of Political Science 57(2): 459–477.

46.

France 24 (2019) Austrian Government Plunged into Crisis Over ‘Ibiza Affair’. Paris, France: France 24. Available at: https://www.france24.com/en/20190518-austrian-government-plunged-crisis-over-ibiza-affair.

47.

Funk

(1996) The impact of scandal on candidate evaluations: an experimental test of the role of candidate traits. Political Behavior 18(1): 1–24.

48.

Galvis

ÁF

Snyder

Jr Song

(2016) Newspaper market structure and behavior: partisan coverage of political scandals in the United States from 1870 to 1910. The Journal of Politics 78(2): 368–381.

49.

Garside

Watt

Pegg

(2016) The panama papers: how the world’s rich and famous hide their money offshore. The Guardian. Available at: https://www.theguardian.com/news/2016/apr/03/the-panama-papers-how-the-worlds-rich-and-famous-hide-their-money-offshore.

50.

Gathmann

(2011) German defense minister accused of plagiarism. Der Spiegel. Available at: https://www.spiegel.de/international/germany/guttenberg-s-printing-press-german-defense-minister-accused-of-plagiarism-a-745891.html.

51.

Golder

(2006) Pre-electoral coalition formation in parliamentary democracies. British Journal of Political Science 36(2): 193–212.

52.

Graham

Haidt

Nosek

(2009) Liberals and conservatives rely on different sets of moral foundations. Journal of Personality and Social Psychology 96(5): 1029–1046.

53.

Groseclose

(2001) A model of candidate location when one candidate has a valence advantage. American Journal of Political Science 45(4): 862–886.

54.

Gutiérrez-Romero

LeBas

(2020) Does electoral violence affect vote choice and willingness to vote? Conjoint analysis of a vignette experiment. Journal of Peace Research 57(1): 77–92.

55.

Hainmueller

Hopkins

(2015) The hidden American immigration consensus: a conjoint analysis of attitudes toward immigrants. American Journal of Political Science 59(3): 529–548.

56.

Hainmueller

Hopkins

Yamamoto

(2013) Causal inference in conjoint analysis: understanding multidimensional choices via stated preference experiments. Political Analysis 22(1): 1–30.

57.

Hankinson

(2018) When do renters behave like homeowners? High rent, price anxiety, and nimbyism. American Political Science Review 112(3): 473–493.

58.

Heath

Karnitschnig

(2019) Austrian vice chancellor Strache resigns over scandal. Politico. Available at: https://www.politico.eu/article/austrian-vice-chancellor-resigns/.

59.

Henley

(2016) Iceland pm steps aside after protests over panama papers revelations. The Guardian. Available at: https://www.theguardian.com/world/2016/apr/05/iceland-prime-minister-resigns-over-panama-papers-revelations.

60.

Hobolt

Leeper

Tilley

(2020) Divided by the vote: affective polarization in the wake of the Brexit referendum. British Journal of Political Science 51(4): 1476–1493. DOI: 10.1017/S0007123420000125.

61.

Huggler

(2019) Austrian far-right leader under criminal investigation over video scandal. UK News. Availabe at: https://www.telegraph.co.uk/news/2019/06/14/austrian-far-right-leader-criminal-investigation-video-scandal/.

62.

Iacono

(2018) Does community social embeddedness promote generalized trust? An experimental test of the spillover effect. Social Science Research 73: 126–145.

63.

Jones

(2019) Austria’s Kurz Forced to Step Down Despite European Elections Win. London, UK: Reuters. Availabe at: https://www.ft.com/content/99f439be-8013-11e9-b592-5fe435b57a3b.

64.

Juanchich

Sirota

Bonnefon

(2020) Anxiety‐induced miscalculations, more than differential inhibition of intuition, explain the gender gap in cognitive reflection. Journal of Behavioral Decision Making 33(4): 427–443. DOI: 10.1002/bdm.2165.

65.

Karasz

(2012) Hungarian president resigns amid plagiarism scandal. New York Times. Available at: https://www.nytimes.com/2012/04/03/world/europe/hungarian-president-pal-schmitt-resigns-amid-plagiarism-scandal.html.

66.

Klüver

Spoon

(2020) Helping or hurting? how governing as a junior coalition partner influences electoral outcomes. The Journal of Politics 82(4): 223. DOI: 10.1086/708239.

67.

Kuhn

Solaz

van Elsas

(2018) Practising what you preach: how cosmopolitanism promotes willingness to redistribute across the European union. Journal of European Public Policy 25(12): 1759–1778.

68.

Kumlin

Esaiasson

(2012) Scandal fatigue? Scandal elections and satisfaction with democracy in Western Europe, 1977-2007. British Journal of Political Science 42(2): 263–282.

69.

Lau

Redlawsk

(2001) Advantages and disadvantages of cognitive heuristics in political decision making. American Journal of Political Science 45(4): 951–971.

70.

Leeper

Hobolt

Tilley

(2020) Measuring subgroup preferences in conjoint experiments. Political Analysis 28(2): 207–221.

71.

Martin

Stevenson

(2001) Government formation in parliamentary democracies. American Journal of Political Science 45(1): 33–50.

72.

Martin

Vanberg

(2005) Coalition policymaking and legislative review. American Political Science Review 99(1): 93–106.

73.

Mondak

(1995) Competence, integrity, and the electoral success of congressional incumbents. The Journal of Politics 57(4): 1043–1069.

74.

Obermaier

Obermayer

(2016) A storm is coming. Panama Papers. Available at: https://panamapapers.sueddeutsche.de/articles/56fec0cda1bb8d3c3495adfc/.

75.

Puglisi

Snyder

Jr (2011) Newspaper coverage of political scandals. The Journal of Politics 73(3): 931–950.

76.

Ramalingam

Morales

Walker

(2018) Varying experimental instructions to improve comprehension: punishment in public goods games. Journal of Behavioral and Experimental Economics 73: 66–73.

77.

Reuters (2019) Austrian vc Caught Discussing Alleged Shady Deals, and Pundits Cry ’Russia Took Over the Country. London, UK: Reuters. Available at: https://www.rt.com/news/459663-russia-pundits-austrian-strache-video/.

78.

Sarmiento-Mirwaldt

Allen

Birch

(2014) No sex scandals please, we’re French: French attitudes towards politician’ public and private conduct. West European Politics 37(5): 867–885.

79.

Saxton

Barnes

(2020) Sex and ideology: liberal and conservative responses to scandal. Journal of Elections, Public Opinion and Parties: 1–12.

80.

Schofield

Laver

(1985) Bargaining theory and portfolio payoffs in european coalition governments 1945-83. British Journal of Political Science 15(2): 143–164.

81.

Sigurdardottir

(2016) Iceland Government Appoints New PM, to Call Early Elections. London, UK: Reuters. Available at: https://www.reuters.com/article/uk-panama-tax-iceland-idUKKCN0X31B3?edition-redirect=uk.

82.

Spoon

J-J

Klüver

(2017) Does anybody notice? How policy positions of coalition parties are perceived by voters. European Journal of Political Research 56(1): 115–132.

83.

Stokes

(1963) Spatial models of party competition. American Political Science Review 57(02): 368–377.

84.

Stone

Simas

(2010) Candidate valence and ideological positions in U.S. house elections. American Journal of Political Science 54(2): 371–388.

85.

Strom

(2000) Delegation and accountability in parliamentary democracies. European Journal of Political Research 37(3): 261–289.

86.

Teele

Kalla

Rosenbluth

(2018) The ties that double bind: social roles and women’s underrepresentation in politics. American Political Science Review 112(3): 525–541.

87.

Thompson

(2013) Political Scandal: Power and Visibility in the Media Age. Hoboken, NJ: John Wiley & Sons.

88.

Verge

(2020) Parliamentary Affairs. Oxford, UK: Oxford University Press. Available at: DOI: 10.1093/pa/gsaa052.Too few, too little: parliaments’ response to sexism and sexual harassment

89.

von Sikorski

(2018) The aftermath of political scandals: a meta-analysis. International Journal of Communication 12: 25.

90.

Sikorski

Heiss

Matthes

(2020) How political scandals affect the electorate. Tracing the eroding and spillover effects of scandals with a panel study. Political Psychology 41(3): 549–568.

91.

Warwick

(1996) Coalition government membership in west European parliamentary democracies. British Journal of Political Science 26: 471–499.

92.

Weise

(2019) Austrian far-right leader filmed offering public contracts for campaign support. Politico. Available at https://www.politico.eu/article/austria-far-right-freedom-strache/.

93.

Wright

Levy

Citrin

(2016) Public attitudes toward immigration policy across the legal/illegal divide: the role of categorical and attribute-based decision-making. Political Behavior 38(1): 229–253.

94.

Zaller

(1992) The Nature and Origins of Mass Opinion. Cambridge, UK: Cambridge University Press.

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

0.92 MB