Jorge Otavio de Souza Silva
Luciano Rossoni
Clayton Pereira Gonçalves- 1Graduate Program in Administration, Grande Rio University, Rio de Janeiro, Brazil
- 2Graduate Program in Administration, University of Brasília, Brasília, Brazil
- 3Department of Administrative and Social Sciences, Três Rios Institute, Federal Rural University of Rio de Janeiro, Três Rios, Brazil
Why do voters vote for candidates known for corruption allegations? We herein suggest two possible explanations for it. First, the choice context affects voters’ risk assessment, since voting for a candidate considers other candidates because aversion to extreme alternatives can change their preferences. Second, the judgment, according to which, candidates’ features are relevant, is not impartial because previously defined preferences set the relevant ones. We also seek to assess such judgments from a moral perspective supported by the understanding that individuals presenting utilitarian ethical orientation substantiate their choices of candidates who deliver results for voters. In contrast, individuals accounting for a deontological orientation focus on means and principles. We designed a vignette experiment with 223 respondents who were randomly exposed to two different scenarios in order to assess such elements empirically: two candidates, one newcomer and one experienced candidate whose accounts were disapproved by the court of auditors; three candidates, the two previous candidates and one experienced candidate who was convicted of corruption. Based on the results, the presence of the most corrupt candidate significantly increased the probability of voters voting for the newcomer; thus, the stronger the utilitarian orientation, the greater the likelihood of voters voting for the experienced candidate, mainly when there is a more corrupt candidate in the race.
1 Introduction
The current political scenario in several countries, including Brazil, reflects critically on elections. On the one hand, the population is apparently less tolerant of corruption and morally reprehensible attitudes, even if they are legal (Arvate and Mittlaender, 2017; Lago, 2019). On the other hand, there is the frequent reelection of politicians involved in corruption or unethical practices, or of their associates (Barros et al., 2019; Boas et al., 2019). The motivations behind each vote may explain this contradiction.
It is possible to compare the electoral environment to markets in general by understanding the essence of the act of voting as an exchange (Philipson and Snyder, 1996), and to use studies related to consumer decision-making to understand voters’ behavior (Gonçalves and Ayrosa, 2022; Philipson and Snyder, 1996). Viewing the individual no longer as purely rational at the time to make choices, but also as a being influenced by subjective factors is increasingly accepted because it introduces new perspectives for the study of voters as consumers (Barros et al., 2012; Whelan et al., 2016).
Even if voters are aware, informed, and confident of their choices, subjective aspects are unlikely to influence their decision. According to studies on decision-making processes, no choice is entirely rational; therefore, when one makes a decision, others often use cognitive shortcuts unconsciously. These shortcuts can help them make the best choice possible with the least effort (Colombo and Steenbergen, 2021). These shortcut types can be the most varying ones, yet, they are related to a series of interferences voters are exposed to. Social, cultural, genetic, and many other influences can build and shape their personality, as well as affect their choices (Kahneman, 2011).
Several studies have set out to analyze electoral actions toward corruption. Vera (2020) explored how candidates’ competence and corruption influenced voters’ willingness to punish corrupt politicians, in Peru. Bøttkjær and Justesen (2021) assessed the role of ‘clientelism’ and partisanship in supporting corrupt politicians in South Africa. Megías et al. (2023) investigated how deontological and consequentialist ethical perspectives shape attitudes toward corruption in Portugal. Breitenstein (2019) analyzed how voters could forgive bribery in exchange for party benefits and economic performance in Spain. Barros et al. (2019) analyzed the impact of ideology on voters’ willingness to vote for corrupt candidates in Brazil. Arvate and Mittlaender (2017) investigated how corruption and inefficiency are punished by voters in an experiment aimed at electoral scenario simulations in Brazil. Finally, Jenke and Huettel (2020) showed how social identity can compete with political considerations at the time to determine voters’ preferences.
The aim of the present study was to find ways to foster discussions about voters’ decision-making process by perceiving them as a type of consumer whose vote is the currency paid in exchange for candidates’ political promises. The main aim of the present research was to investigate the likely association between individuals’ deontological or utilitarian ethical structure and the greater or lesser tolerance for candidates’ political corruption based on a randomized vignette experiment conducted to simulate options for choosing mayoral candidates.
It is important highlighting that the discussion about corruption demands to mention ethics. After all, corruption is found where one lacks ethics (Stukart, 2003). Conceptualizing ethics is not an easy task, because, despite the centuries of discussions on the subject, there is no single definition accepted by the entire society (Conway and Gawronski, 2013). We sought to introduce the two most accepted and frequently discussed classifications linked to modern ethics, namely: deontology and utilitarianism. Overall, the first values focus on faithful and strict compliance with rules and moral principles, whereas the second ones focus on the consequences of a given action (Gawronski and Beer, 2017; Nozick, 1981). Likewise, the definition of corruption is not undisputed because this term encompasses a whole series of attitudes, such as, theft, fraud, bribery and nepotism, including everything from minor behavioral deviations to highly severe and complex profile (Brei, 1996; Rose, 2018). Corruption will be that practiced by politicians and, therefore, related to public affairs and to the community.
The current study used two distinct scenarios, both addressing the same context, to examine how certain subjective aspects influence voters’ decision-making. These scenarios involved an electoral dispute between mayoral candidates in a fictitious Brazilian municipality and were based on a randomized vignette experiment (Boas et al., 2019; Carreras and Vera, 2018; Vera, 2020). Three candidate profiles were presented: the first candidate was honest but politically inexperienced, running for office for the first time; the second candidate was an incumbent seeking re-election who, despite achieving satisfactory results in public management, had his accounts rejected by the State Audit Court; and the third candidate, also seeking re-election, had been tried and convicted of corruption. In the first scenario, only the first two candidates were presented, whereas the second scenario included the third candidate with a history of corruption. Participants were randomly assigned one of these scenarios and asked to vote for one of the presented candidates.
Unlike other studies that have ignored the real conditions for choosing one of the candidates (Boas et al., 2019; Weitz-Shapiro and Winters, 2017), our intention was to simulate real voting situations, which allowed us to hypothesize that the presence of a third, and more corrupt candidate, would trigger the aversion bias toward extremes (Tversky, 1994). This process, which could lead to more choices for the less corrupt newcomer candidate, has shown that the mere presence of a third candidate, who was more corrupt than the others, could influence voters’ voting intentions; yet, there are voters who were lenient toward corruption if there was some sort of beneficial counterpart for their community – this find was already pointed out in previous studies (Bonifácio, 2013).
We assessed voters’ ethical orientation through the Survey of Ethical Theoretic Aptitudes (SETA), which was applied to classify voters into individuals presenting a deontological or a utilitarian profile. In order to do so, we hypothesized that the more utilitarian the individual, the greater its likelihood for choosing the mayoral candidate known for history of achievements, despite the corruption allegations. In addition, we also hypothesized that the presence of a third (more corrupt) candidate led to an even higher utilitarianism, because it triggered the aversion to extremes, to a greater extent. The findings can be seen as advancements in the studies by Boas et al. (2019) and Weitz-Shapiro and Winters (2017), as they took into account factors intrinsic to voters’ moral judgment when they punish corrupt politicians through their voting. This attitude is in compliance with a research agenda supported by recent changes in elections worldwide (Lago, 2019).
2 Framework
2.1 Elections as markets
In a broader sense, markets can be understood as environment where individuals provide currency in exchange for a given product or service (Philipson and Snyder, 1996), where exchanges of goods or services offered by producers and demanded by buyers are carried out (Baker, 2007). This environment embodies social structures characterized by close social associations among firms, workers, suppliers, customers and governments (Fligstein and Dauter, 2007). The common idea lies on the understanding that a place, which is not necessarily physical, where potential consumers of any good, service or idea, and where sellers of all good types, interact to satisfy their respective desires and needs.
Thus, giving specific meaning to the sense of market allows associating it with the electoral process (Philipson and Snyder, 1996; Schumpeter, 1942). Both the product market and the election market regard transactions, perception of values, product differentiation and marketing campaigns. While consumers offer money in exchange for goods or services, voters give their vote to the candidate whose speech they believe is the best in public administration terms (Gonçalves and Ayrosa, 2022). Salespeople like political candidates, in their turn, account for persuading and convincing others to reach and win over their consumer target (Barros et al., 2019). So much so, that the analogy between voter/consumer and candidate/product is common in the literature, according to which, candidates seek to sell their ideas and proposals to voters who likely want to get them (Perannagari and Chakrabarti, 2020; Scammell, 1999). Based on the perception that voters behave as consumers, it is possible associating some studies on consumers with voters. The focus of the current study is to assess voters’ decision-making process when they must decide who to vote for, mainly when it comes to their biases, as well as to the effects of ethical orientation.
Simon (1955, 1956) was one of the first ones to perceive and theorize about the influence of behavioral biases by taking into account decision-making behaviors. The behavioral approaches to markets, including the electoral ones, gained more prominence at late 20th century through the work by Kahneman and Tversky. They were recently boosted by the work by Richard Thaler. Their main reflection on this approach lies on the understanding that economic agents have cognitive limitations in decision-making processes, and it would prevent them from acting as suggested by neoclassical economists, i.e., in a purely rational way. It happens because it would be impossible to process all information available about likely alternatives, as well as about their respective consequences (Simon, 1955, 1956). These limitations can be both conscious and unconscious. In other words, economic agents can, intentionally or not, choose the non-optimal alternative. Vera (2020) showed how the perception of candidates’ competence in electoral contexts and the prevalence of corruption influence voters’ willingness to punish corrupt politicians. However, the research also identified that voters sometimes tend to accept corruption from candidates considered competent, mainly when these candidates deliver public works that meet voters’ needs.
Some reflections on voters’ decision-making are enough to question their behavior, as suggested by the theory of economic rationality (Colombo and Steenbergen, 2021). It is up to voters to decide if it is worth investing time and effort in searching for data and information about the candidates if they are aware of their votes’ marginal effect, which could change the final result of an election (Downs, 1957). Thus, if voters are aware that their decision is virtually insignificant, they would seek to measure the cost of their vote for themselves. Furthermore, there is individuals’ natural difficulty in choosing something whose result, in addition to not being immediate, cannot be seen, only perceived.
Thus, criticism emerges from the so-called homo economicus, who is perfectly rational and whose choices are perfect. They believe that subjective aspects exert little, or no, influence on their decision-making process, to the point of not being taken into consideration. Economic agents, who are herein referred to as voters, naturally have some kind of pre-established label set by the social environment they are inserted in. Therefore, when they are faced with a decision, no matter how simple it might be, it is difficult to ignore factors such as social attributes, cultural identity, ways of thinking, acting, and worldview, which can influence their choices (Stoetzer and Zittlau, 2020). As highlighted by Jenke and Huettel (2020), the identity of a social group can compete with political and ethical considerations at the time to form electoral preferences, and it can decisively influence voters’ behavior.
2.2 Utilitarian and deontological ethical orientation
One of these subjective aspects, the so-called modern ethics, despite centuries of discussion on it, basically revolves around two major schools of thought: the deontological and utilitarian thinking. This process does not mean lack of alternative approaches, but that the deontological/utilitarian dichotomy has been the basis for discussions on this topic (Brady and Wheeler, 1996). Nozick (1981) stated that much of the debate on ethics could be classified into one of these two categories.
According to the deontological ethics, the moral value of a given action lies on the action itself. Good individuals are those who act correctly; who obey the norms, rules and moral principles, regardless of their effects (Brady, 1990). Thus, they are those who perform a correct action, who obey principles, act well, even if the result is bad. Similarly, those who behave badly, even if the consequences of their action are beneficial, act badly. As discussed by Megías et al. (2023), these ethical guidelines have straight influence over the way voters respond to corruption. Therefore, the origin of good would be the act of acting, the action itself, of doing the right thing. Deontological individuals are those motivated by faithful compliance with rules, principles or with any other type of formal feature. Therefore, those who do not comply with them act badly, even if they lead to good outcomes (Dias and Bento, 2011).
The utilitarian ethics, in its turn, focuses on the effects generated by a given attitude, i.e., the attitude itself has no value (Brady, 1990). From a utilitarian perspective, an action is good simply if its result is good. Similarly, an action is bad if it leads to bad consequences. In this case, the attitude itself, or the means used to achieve a given end, are of little importance; it is enough to simply observe its effects (Brady and Wheeler, 1996; Costa, 2002). Therefore, utilitarian scholars understand that decisions should be assessed as morally correct or incorrect based on the analysis of results generated by them (Conway and Gawronski, 2013). A simple example can illustrate the difference between these two schools of thought: a poor and needy individual who steals food from a certain establishment to feed itself should be condemned. According to deontologists, because this individual performed the act of stealing, it is bad, itself. Based on utilitarian beliefs, the consequence of such an attitude is good, justifiable; therefore, the individual acted correctly (Megías et al., 2023).
It is worth noticing that the herein introduced concepts of deontology and utilitarianism bring along other connotations, but this discussion is not the aim of this study. Deontology is associated with expressions, such as “Kantian ethics” and “formalism,” whereas utilitarianism is often related to “consequentialism,” “teleology,” “pragmatism,” among other terms. The definitions of deontology and utilitarianism simply aim at representing a set of ideas whose central position concerns obedience and compliance with rules, respectively, and the assessment of consequences of a given decision (Brady and Wheeler, 1996; Megías et al., 2023).
There are several ways to assess an individual’s ethical orientation. Oftentimes, it is possible to perceive the way a person thinks and sees society, in general, depending on the answers given by it to certain questions; thus, individuals can be classified based on a certain ethical profile (Megías et al., 2023). The Survey of Ethical Theoretic Aptitudes (SETA) (Brady, 1990) was found, among the validated scales, to be the one best matching and contributing to the aims of the current research. In addition to being widely accepted and to having already been used in several studies (Kumar and Strandholm, 2002; Love et al., 2020; Middleton and Byus, 2004; Smith and Pulver, 2009), it addresses the two dominant ethical currents, which are the herein adopted ones, namely: deontologism and utilitarianism.
2.3 Corruption and elections
The debate on the concept of corruption is equally historical, dynamic and has changed, overtime, due to different social and cultural contexts (De Vries and Solaz, 2017). Although there is certain consensus on its association with the violation of rules in order to obtain advantages, to this day, there is no single, peaceful and unquestionably meaning of it acceptable by society. Furthermore, the fact that several different groups and classes coexist in the same community contributes to the hard time reaching consensus on what is corrupt, or not. Depending on the context, these groups and classes can interpret the idea of corruption in different ways (Barros et al., 2019; Breitenstein, 2019).
Nevertheless, corruption is currently linked to the political environment and involves public agents and the mismanagement of social and collective resources. Corruption was herein considered as the undue transfer of what is public to the private sector (Shleifer and Vishny, 1993), encompassing illegal payments – financial or otherwise – to obtain, accelerate, or prevent the performance of a service by a public or private employee. Miranda (2018) offers a valuable attempt to unify the concept of corruption within political science, framing it as a comprehensive category that encompasses personal or political gain obtained through illegal payment (financial or otherwise). This difficulty is further complicated by the variety of corrupt practices (Brei, 1996). Arvate and Mittlaender (2017) demonstrated how the perception and punishment of corruption can change depending on the political context and government efficiency. However, the essence remains in the misconduct of public agents pursuing goals that conflict with collective interests.
Recent studies have significantly contributed to the debate on the context of corruption and candidate selection by analyzing how factors, such as candidate competence, ideology, and government efficiency, influence voters’ responses (Arvate and Mittlaender, 2017; Barros et al., 2019; Bøttkjær and Justesen, 2021; Breitenstein, 2019; Megías et al., 2023; Vera, 2020). These studies have highlighted that corruption perception and punishment significantly change depending on political context and governmental efficiency.
According to Boas et al. (2019), norms, laws or any other anti-corruption rule type do not have impact on voters’ decision-making process, no matter how strong and efficient they may seem. Thus, information about the approval or rejection of a given candidate’s accounts, according to this author, does not translate into consequences at the polls. The population’s concern with more concrete and tangible issues, such as employment and health, and their impact on each voter’s voting intention, was also brought to light by him. De Vries and Solaz (2017) also sought to understand voters’ decision-making process when candidates present corruption elements. According to them, group identity has strong influence on voting intention. Their findings suggest that voters can support corrupt candidates as long as they share a group identity with them. This outcome corroborates other studies carried out in Brazil (Barros et al., 2019; Boas et al., 2019; Bonifácio, 2013), according to which, voters tend to minimize corruption aspects when they ideologically identify themselves with corrupt candidates. However, given the corruption scandals Brazil has witnessed in the last 10 years, identifying a candidate as corrupt may have greater impact on voting intentions for mayoral candidates in municipal elections, since it has gained considerable prominence in the electoral debate (Agerberg, 2020; Bélanger and Meguid, 2008).
3 Hypotheses
Three hypotheses were advocated for in the present study. The first one assumes that the mere presence of a candidate who is noticeably more corrupt than the others can change the votes’ configuration in times of choosing mayoral candidates. It provides voters with a new perspective toward the two candidates who were already in the race. In other words, the division of votes between two candidates would be one, but the addition of a third candidate, even if he is overlooked in comparison to the others, would change the division of votes for each of the other two candidates. Clearly, there is no reason for the third candidate to gain votes, since there is another less-corrupt competitor in the race, who is equally efficient in delivering results.
The presence of a candidate who is more corrupt than the others would be justified by the idea that voters would be able to better compare and confront the candidates based on a common feature (to be corrupt), which could influence their voting intention (Agerberg, 2020; Breitenstein, 2019; Vera, 2020). The idea of aversion to extremes could be configured within this context.
According to the concept introduced by Tversky (1994), a new option should not, from a rational viewpoint, change the preference level between the original alternatives. In other words, a disregarded option could not become the preferred option when new alternatives are added. However, this author stated that the simple presence of new options can twist or change individuals’ preferences. The aversion to extremes would be the reflection of the principle of loss aversion. Thus, the idea that losses have more negative effect than proportional gains leads decision makers to avoid extreme positions and to opt for central or morally desired alternatives (Simonson and Tversky, 1992).
Voters compare a candidate to other candidates in the race when they make a decision about it, and this process changes their perception and judgment about the other candidates (Agerberg, 2020; Breitenstein, 2019; Simonson and Tversky, 1992; Tversky, 1994; Vera, 2020). Thus, a context with two candidates presents a comparison and judgment outline between these two, only. The context of perception and judgment changes as a third candidate is added to the race, even more so because it is considered more corrupt than the others. Furthermore, the aversion to extremes (Simonson and Tversky, 1992; Tversky, 1994), applied to the most corrupt one, must change voters’ reference point, so that those who use corruption as reference for judgment will have their decisions affected by the presence of a third, more corrupt, candidate, because they have an aversion to loss.
Thus, if one has in mind a choice context comprising two hypothetical mayoral candidates: candidate A, a political newcomer; and candidate B, an experienced candidate with failed accounts, one can assume a given decision, which would be the median indicator of illicit activity. However, the inclusion of a third, more corrupt, candidate, who was convicted of embezzlement, would change the context and trigger the aversion to extremes bias (Tversky, 1994). Preferences are often already established in voting contexts (Alesina and Passarelli, 2019); in the case of the current study, these preferences lead to greater acceptance of candidates without any evidence of corruption (Agerberg, 2020). The aversion to extremes bias would lead to choosing the candidate presenting morally desirable features (Wang and Chen, 2019); in our case, the newcomer candidate. Therefore, the first hypothesis of the current study assumes that:
H1: The presence of a third, more corrupt, candidate in the race, increases the intention to vote for the new candidate, who does not have any corruption allegation.
The second hypothesis assumes that individuals’ ethical orientation can affect their voting intention. According to Piurko et al. (2011), voters’ personal values, such as hedonism, benevolence and attachment to traditions, are decisive to their political choices. Thus, political decisions would lie on individuals’ intuitive field rather than on their rational field (Smith et al., 2017). Megías et al. (2023) highlighted how deontological and consequentialist ethical perspectives have straight influence on individuals’ attitudes toward corruption, and it suggests that these ethical frameworks can determine whether a voter accepts or rejects corrupt candidates. Therefore, from the perspective of individuals’ ethical structure as subjective and intrinsic aspect of their personality, it can likely influence their decision-making process.
Assumingly, the more utilitarian the individual, the less likely it is to vote for an honest politically inexperienced mayoral candidate. Similarly, the greater the deontological ethical structure of an individual, the greater its likelihood for voting for a non-corrupt candidate. It is so, because an individual’s ethical profile can likely lead it to vote, or not, for a dishonest candidate.
More specifically, it is assumed that a utilitarian individual would consider voting for a corrupt candidate, as long as it is politically efficient, in other words, if it is competent enough to present public improvements, be them social, economic, cultural, among others (Breitenstein, 2019; Vera, 2020). This voter type (utilitarian ethical profile) would focus on results a given candidate could deliver, it would not care about how it would achieve them (Costa, 2002; Megías et al., 2023.
Carreirão (2008) described the profile of politicians commonly portrayed in Brazil through expressions, such as “he steals, but he gets things done,” “it makes no difference whether a politician steals, or not, what matters is that it does things the population needs” or “it is better to have a politician who works a lot, even if it steals a little, than a politician who does little work and steals nothing.” This tolerance for corruption, mainly that associated with results perceived as beneficial, was also explored by Vera (2020), who highlighted how the perceived competence of a candidate can influence voters’ willingness to accept corruption. Similarly, it is possible assuming a voter with deontological ethical orientation, i.e., the one who values faithful compliance with rules and norms would never vote for a corrupt candidate, regardless of the results it is capable to deliver (Dias and Bento, 2011). Even if the candidate is politically efficient and achieves significant results for the community, assumingly, voters with deontological profile will reject it if it used unethical means to achieve them. Thus, the second hypothesis considers that:
H2: The more utilitarian the individual, the lesser likely it is to vote for a politically inexperienced and honest candidate.
The third hypothesis assumes that the voter’s ethical profile intensity can also influence their voting in different choice contexts. In other words, the individual with utilitarian structure will be more likely to vote for a candidate identified with intermediate corruption level in the presence of a visibly more corrupt candidate, but who delivers results and whose accounts have been rejected, which suggests a kind of decoy effect (Huber et al., 1982). The opposite is also true, if there is a candidate who is noticeably more corrupt than the others, a voter with deontological profile would be more likely to vote for the inexperienced and honest candidate in the scenario with only two candidates (Agerberg, 2020).
However, the aversion to extremes context is only possible in scenarios with three or more options, which provides a figure of extreme, according to which, one of the options appears as neutral reference that corresponds to the status quo (Kahneman et al., 1991; Simonson and Tversky, 1992). Furthermore, presenting two different scenarios to the same respondents could lead to results’ contamination; therefore, we sought to compare the ethical orientation effect on voting for the newcomer candidate over the experienced one in two scenarios: control scenario, with two candidates; treatment scenario, with three candidates.
Accordingly, if one has in mind the idea of aversion to extremes, the presence of the third more corrupt candidate in the scenario with three candidates would be interpreted as risk to the election of a less corrupt candidate; thus, the disadvantages attributed to the option rejected by the voter stands out (Bélanger and Meguid, 2008) in comparison to advantages offered by the candidate of their choice (Simonson and Tversky, 1992; Tversky, 1994). Loss aversion will cause some voters to choose the option that reinforces their beliefs (Resende et al., 2024), including the ethical ones, or that representing the status quo (Yair et al., 2020). Thus, individuals tend to be more conservative in their voting choices in uncertainty contexts or at risk of unsatisfactory results, since the disadvantages of change would seem more important than their benefits (Alesina and Passarelli, 2019; Kahneman et al., 1991). There is higher trend for choosing a mayoral candidate based on personal preferences, such as ethical orientation, within uncertainty contexts (Megías et al., 2023; Resende et al., 2024).
Actually, we understand that ethical orientation acts as pendulum between choices for the newcomer and the less corrupt experienced candidates (Agerberg, 2020; Resende et al., 2024). Change in the probability of voting changes depending on ethical orientation, so that the more utilitarian the voter, the more likely for it to vote for the experienced candidate, with disapproved accounts in case there is a third, more corrupt, candidate in the race. The opposite is also true, the presence of a third candidate will make a deontologically oriented voter more likely to vote for the candidate of its preference, the newcomer one. According to the third hypothesis, we can assume that:
H3: If there is a candidate more corrupt than the others in the race, then, the more utilitarian the voter, the lesser likely it will be to vote for the honest politically inexperienced candidate, i.e., similarly, the greater the probability of voting for the more experienced candidate, who delivered results, despite the fact that its accounts were rejected.
4 Research design
At the same time, to assess the effect of choice context and ethical orientation on the choice for mayoral candidates, we conducted an experimental study using randomized vignettes. According to Atzmüller and Steiner (2010, p. 128), ‘a vignette is a short, carefully constructed description of a person, object or situation, that represents a systematic combination of features.’ Similar to other vignette studies conducted through experimental surveys (Boas et al., 2019; Carreras and Vera, 2018; Gaines et al., 2007; Vera, 2020), our study consisted of two elements: vignettes – where each respondent was randomly assigned to a single scenario (between-subjects design) composed of two or three hypothetical descriptions of mayoral candidates; and a traditional survey – where all respondents answered the same set of questions, including our independent variable and covariates used to control for confounding variables. It is important to clarify that while the presentation order of the vignettes was randomized among respondents, the selection of participants was not, as responses were obtained through voluntary participation.
4.1 Data collection
An electronic form was applied for data collection purposes, the link to the form was sent by email to a list acquired from telephone and email contact services. Two thousand of the 16,537 individuals who were invited to participate viewed the questionnaire (12.1% of the total); 381 of them (19.1%) filled it up, 245 (64.3%) of them fully responded it and 136 (35.7%) did not fully respond it. An attention check question was inserted in it: of the 245 individuals who fully responded to the questionnaire only 6 were excluded from the study because they did not meet the attention check variable (Berinsky et al., 2014). Then, we ruled out 5 respondents who voted for candidate C. Finally, we removed 11 extreme outliers (standard error > 3), which provided us with a valid sample of 223 cases.
4.2 Vignettes and dependent variable
Vignettes used by Boas et al. (2019), who, in their turn, replicated the vignettes by Weitz-Shapiro and Winters (2017), were adjusted to contrast the choice between the newcomer candidate and candidates accused, or convicted, of corruption. These vignettes are described in Table 1.
Table 1. Scenarios and vignettes used in the experiment.
The Scenario with 2 candidates was chosen as control group, and part of the randomly defined respondents had to choose between the two mayoral candidates: newcomer (A), with no political history, but who represents a renewal in politics; or experienced candidate (B), whose accounts were rejected by the Court of Auditors. The Scenario with 3 candidates was adopted as treatment group; its respondents were also randomly defined and had to choose one of the three candidates: candidates A or B, who were equally exposed just as in the control group; or candidate C, the extreme case, who presented the same features observed in candidate B, but who was convicted of embezzlement.
Thus, our experiment followed a between-subjects design, since respondents were exposed only to one scenario. In operational terms, variable treatment was manipulated as dummy variable: 0 for the scenario with 2 candidates, 1 for the scenario with 3 candidates. The dependent variable was also defined as dummy variable: 1 for those who chose candidate A, the newcomer; 0 for those who chose candidate B, the experienced candidate with failed accounts. Respondents who chose candidate C, who was the extreme option, were discarded, because they were rare cases, not to mention that it was counterintuitive (Table 2).
Table 2. Descriptive statistics applied to the two scenarios.
It is important pointing out that, unlike Boas et al. (2019), and Weitz-Shapiro and Winters (2017), who used a four-item scale to assess the probability to vote for each candidate, we chose to force dichotomous choices by comparing alternatives side by side. We did so because, as Hainmueller et al. (2015) point out, comparative vignette experiments have greater external validity and are closer to real decisions. Secondly, such a design was in compliance with the argument that the choice context represented by different mayoral candidate alternatives changes the value of these very alternatives (Simonson and Tversky, 1992; Tversky and Simonson, 1993; Tversky, 1994).
4.3 Further measures
Ethical Orientation. Right after choosing the candidates, respondents were asked to answer the Survey of Ethical Theoretic Aptitudes (SETA) (Brady, 1990). The scale was originally developed in English and, later on, translated and tested before being used. It was done to ensure its applicability to the current participants, and it enhanced its validity. SETA consists of 15 two-alternative questions aimed at assessing these questions together to capture whether individuals had a more deontological or utilitarian ethical orientation. After weighing each of the alternatives, respondents were classified as follows: +7, +6, +5: “extremely utilitarian”; +4, +3, +2: “slightly utilitarian”; +1, 0, −1: “neutral”; −2, −3, −4: “slightly deontological”; −5, −6, −7: “extremely deontological.” We used these scores to operationalize independent variable ‘ethical orientation’ to create a 5-point scale, according to which, the lower the value, the more deontological; and the higher the value, the more utilitarian. We also kept the variable with natural scores to check results’ robustness, which were convergent in all cases.
Control variables. In addition to vignettes and ethical guidance, some control variables were also collected in order to mitigate their confounding effects, just as observed in previous studies (Boas et al., 2019; Weitz-Shapiro and Winters, 2017). Yet, these variables were adopted to assess whether control and treatment groups were equal to each other, as well as whether individuals ideologically identified themselves with the left or the right wing based on a 10-point scale, according to which, the higher the value, the more to the right wing. It was also checked whether respondents were aware of any corruption case in the last few years in their municipality. Schooling was analyzed by ordering it into six levels, from elementary school to PhD degree. Income was assessed based on 7 categories that ranged from less than one minimum wage to more than 20 minimum wages. The minimum wage in Brazil in 2024 is 1,412.00 Brazilian reais, which is equivalent to 246.00 US dollars. It is important to note that the average household income in Brazil is 1,848.00 reais (U$ 322.00), according to the The Brazilian Institute of Geography and Statistics (IBGE).1 More complex municipalities are more challenging when it comes to prior experience; therefore, the size of respondents’ municipality was split into 5 levels ranging from up to 50,000 inhabitants, from 50,000 to 100,000 inhabitants, from 100,000 to 300,000 inhabitants, from 300,000 to 1 million inhabitants, and more than 1 million inhabitants. We should emphasize that we did not use the city classification scale from Brazil’s official statistical agency (IBGE), as it only includes three levels with very large differences between the categories. Finally, respondents’ age was collected.
Although the assignment was random, there was a small difference in the number of cases between the two groups. However, according to Table 1, number of inhabitants was an exception to it. All other control variables, including the independent variable (SETA), recorded quite the same means. The between-subjects experimental design assumes sample homogeneity; therefore, the effect of confounding variables on control/treatment groups’ differentiation was assessed through Factorial ANOVA test. There was no significant variable and the models’ R2 was extremely low (0.058). This finding highlighted that the sample in the two groups were homogeneous, and it ensured their comparability.
4.4 Method
Logistic regression models were applied to test the three hypotheses because the dependent variable/candidate A or B choice was binary. Thus, seven models were run, one with the control variables, alone; and the other two were applied to each hypothesis. The treatment variable (Scenario) and independent variable (Ethical Orientation) effect on the odd-numbered models in Table 3 was contrasted to the control variables in order to control their effects. With respect to hypothesis 3, it was tested whether the scenario has mitigated the ethical orientation effect by interaction with the variables. These procedures allowed reducing the effects of any bias on the control and treatment groups’ definition, as well as isolating the effects of other confounding factors (Gaines et al., 2007).
Table 3. Logistic regression coefficients recorded for choosing candidate A (newcomer).
5 Results
Figure 1 provides the percentage of votes for each candidate in the scenario with 2 candidates and in that with three candidates. In total, 62% of respondents chose candidate A in the scenario with two candidates, who was the hypothetically honest newcomer candidate. The choice for candidate A increased to 81% in the scenario with 3 candidates, which included a third candidate whose profile was quite similar to that of candidate B, but with worse features (tried and convicted of embezzlement). This finding corroborates the argument that the aversion to the third candidate, who is clearly corrupt, changed the perception about the other candidates, which led voters to choose A (p = 0.002, Fisher’s exact test). Models 2 and 3 in Table 3 were adopted to test the effect of including a third candidate on the probability of choosing candidate A and to make the results more robust. If one has in mind the model based on control variables, it is possible observing that the change to a scenario with three candidates significantly increased the chance (by 22%) of voting for candidate A (β = 1.018, p < 0.001). This outcome confirmed hypothesis 1.
Figure 1. Choosing mayoral candidates in scenarios with 2 and 3 candidates.
Models 4 and 5, still in Table 3, were adopted to assess the effect of respondents’ ethical orientation on the probability of choosing the newcomer candidate A. According to the results, the higher the SETA scale value, the more utilitarian the respondent and the lower the probability of voting for candidate A (β = −0.758, p < 0.001). This finding confirmed hypothesis 2. In terms of probability recorded for the full sample, it was observed that an extremely deontological and formalist respondent would have 91% chance of choosing candidate A. Sample average (SETA = 2.62) showed 77% probability of choosing candidate A. The probability of totally utilitarian respondents choosing candidate A is 40%. In other words, extremely utilitarian individuals have a 60% chance of choosing candidate B, who is more experienced but has a record of failed accounts.
SETA scale was interacted to the scenario in models 6 and 7 to analyze whether the effect of ethical orientation on candidate choice changes depending on the two scenarios (SETA × Scenario variable). The interaction terms were significant in both models. Results recorded for the coefficient of the control-variables model (β = −1.269, p < 0.05) pointed out that the presence of a third, more corrupt, candidate changed the association between ethical orientation and choice for the new candidate (A). With respect to this effect’s orientation, the scenario with 3 candidates evidenced that the more utilitarian the respondents’ ethics, the lower their probability to vote for candidate A (the new one); consequently, it implies a higher probability to vote for candidate B. This finding confirms hypothesis 3.
Figure 2 depicts the probability of voting for candidate A and illustrates the effect of ethical orientation on both scenarios. According to the scenario with two candidates, the probability of voting for candidate A reaches 86% among extremely deontological individuals (core 1 in the SETA scale) and 57% among extremely utilitarian individuals (score 5 in the SETA). This effect is stronger in the scenario with three candidates, and the probability to vote for candidate A can range from 64% among extremely deontological respondents to 0.2% among the extremely utilitarian ones. In other words, according to the scenario with 3 candidates, the more utilitarian the respondent, the more corrupt the candidate and the higher the probability of voting for candidate B, who is experienced but known for failed accounts.
Figure 2. Probability to vote for candidate A (newcomer) in both scenarios.
Finally, it is essential highlighting the effect of control variables that recorded significant effect on the choice for candidates (Model 1, Table 3). The more politically aligned to the right the voter is, the higher its propensity to vote for candidate A (β = 0.189, p < 0.001). Older age is also related to greater chance to vote for the same candidate (β = 0.034, p < 0.001). It is also interesting noticing that the larger the city the respondent lives in, the lower its probability to vote for candidate A (the newcomer) (β = −0.445, p < 0.001) and to vote for candidate B, who is more experienced but known for failed accounts.
6 Discussion
The aim of the present research was to investigate how the context where mayoral candidates are chosen and voters’ utilitarian and deontological ethical orientation influence the choice made between newcomer candidates and those known for corruption allegations. Thus, these effects were assessed through a randomized vignette experiment to simulate options for choosing mayoral candidates based on two different scenarios, one with only two candidates and another with three candidates (including a third experienced candidate known for corruption allegations).
The electoral environment was compared to a market type, according to which, votes are a sort of currency exchanged between voters and candidates (Gonçalves and Ayrosa, 2022; Whelan et al., 2016). Candidates act as firms whose goal is to persuade and convince buyers that their product stands out from the others. Voters, as consumers, account for choosing the product (candidate) they believe best meets their needs based on their ethical orientation (Resende et al., 2024).
The idea was to find out what interference type it can exert on decision about what candidate to vote for when the voter is exposed to new candidates and to those with corruption allegations by taking each individual’s ethical orientation as factor inherent to its respective personality. Thus, the first hypothesis was to investigate whether the presence of a third candidate in the race (more corrupt than the others) would change voters’ voting intentions based on the scenario with two candidates by using the corruption factor as differentiating element between candidates (Agerberg, 2020; Arvate and Mittlaender, 2017; Bélanger and Meguid, 2008). The current results corroborate previous studies (Vera, 2020) and pointed out that the presence of a candidate who is visibly more corrupt than the others can change voters’ voting intentions. This finding is in compliance with the so-called aversion to extremes effect (Kahneman et al., 1991; Simonson and Tversky, 1992), according to which, the presence of a third option, even of a worse one, can change the choice pattern in comparison to a scenario where voters only have two options. This finding reflects the findings by Agerberg (2020) who observed that voters tend to punish corrupt candidates whenever there is another honest or less corrupt candidate.
The second hypothesis aimed at assessing whether the higher the utilitarianism level, the lower the probability of voting for an honest politically inexperienced candidate rather than to an experienced candidate with corruption allegations. Results have shown that voters with stronger utilitarian orientation tend to avoid voting for less experienced candidates, even for those without history of corruption. Furthermore, these voters rather vote for a candidate who has already shown its ability to promote gains for society, even if it is known for corruption allegations. Some voters become lenient with corruption, if it would bring some sort of beneficial counterpart for the community (Arvate and Mittlaender, 2017), mainly when it comes to utilitarian-orientation voters (Conway and Gawronski, 2013). These findings are similar to those in previous studies (Breitenstein, 2019; Vera, 2020), according to which, corrupt candidates can be voted within certain electoral contexts, mainly when they are well analyzed by voters. Furthermore, these findings are in compliance with the research by Megías et al. (2023), who pointed out that deontological and utilitarian perspectives shape voters’ attitudes toward corruption.
Finally, to test the third hypothesis, it was assessed whether the utilitarianism effect is more influential for the probability of voting for the experienced candidate when a third candidate, who is also experienced but more corrupt, is included in the race. Results highlighted that the presence of a candidate who is notoriously more corrupt makes more utilitarian voters more likely to vote for the experienced candidate, to the detriment of the candidate who has never presented gains for society. In other words, based on this scenario, the more utilitarian the voter, the higher its probability to vote for the candidate known in Brazil as the one who “steals, but gets things done” (Carreirão, 2008). The mere presence of a corrupt candidate can change votes’ configuration in an election, even if no one votes for this candidate. Results pinpoint the intense presence of both the status quo bias and loss aversion (Bélanger and Meguid, 2008) among utilitarian voters, since they prefer and reinforce the maintenance of current situations, which seems to pose fewer risks, to the detriment of changes. The operating bias is pure aversion to extremes when voters are deontologically oriented, and it makes them opt for the newcomer. In other words, voters with deontological orientation meet the results observed in Agerberg (2020) study, according to which, voters tend to punish corrupt candidates and to choose the honest ones.
Thus, it is possible stating that voters’ ethical orientation influences their decision-making about voting, because different results were herein observed depending on their ethical orientation. Utilitarian voters, in their turn, tend to vote for corrupt candidates, whereas deontological voters reject them. These conclusions foster the Brazilian political-social debate, which is visibly focused on corruption issues. The contradiction between the practically unanimous social discourse on fighting corruption and the compromise with it, in certain contexts, is real and triggers reflections (Barros et al., 2019; Boas et al., 2019).
6.1 Contributions
Results in the present study provide arguments that can help developing the intense political and social debate currently taking place in Brazil, as well as studies on elections and political marketing. This topic is of interest to society, in general; therefore, other groups can also, directly or indirectly, benefit from this research.
Furthermore, the current results corroborate the idea of limited rationality proposed by Simon back in the 1950s. Kahneman, Tversky and Thaler stand out as its greatest exponents in present times. According to the collected data, individuals’ ethical profile is a moral and subjective factor; therefore, it is far from their pure and conscious reason, and influences their voting intention (Resende et al., 2024). It is important highlighting that tolerance toward corruption in the current Brazilian political scenario and in the discourse of a large part of society has been increasingly lower, regardless of ideologies and political-partisan positions. It is one of the most prominent topics in current debates (Yair et al., 2020).
Results in the present research also provide important considerations for the fighting-corruption debate, which has been so popular in Brazil, in recent years. Knowing that illicit and/or immoral practices exist and are tolerated by the population, depending on the context, provides an important perspective on the ways to fight it. Furthermore, from a marketing perspective, this finding adds ingredients assumingly to have been previously imperceptible, and it can now be adopted as strategy aimed at political campaigns. Knowing that a certain group of voters, under certain circumstances, can tolerate corrupt practices opens room for new challenges for politicians and, most of all, for society, in general.
Finally, it is necessary emphasizing that, unlike Boas et al. (2019) and Weitz-Shapiro and Winters (2017), who adopted a four-item scale to measure the probability to vote for each candidate, without comparing the probability between different vignettes, option was made for a dichotomous choice format to make straight comparisons between alternatives, since vignette experiments contrast the probability of choosing for experienced but corrupt candidates to the detriment of a honest one when the race holds three or two candidates. This methodological choice presents greater external validity and more closely resembles real decisions in majoritarian elections. It is even useful for interpreting Brazilian elections’ scenarios held in cities with more than 200 thousand inhabitants (two-round elections). The second round always presents two candidates and neither of them reached 50% of the valid votes in the first round. This design is in compliance with the argument, according to which, the context of choice (represented by the presence of different mayoral candidates) changes alternatives’ values (Simonson and Tversky, 1992; Tversky and Simonson, 1993; Tversky, 1994).
6.2 Limitations and suggestions for future studies
The present research has limitations linked to its results, because participants’ income and schooling, which are important socioeconomic indicators, were higher than the national average. Assumingly, these results cannot be extrapolated to all individuals with active electoral capacity in the country. The turbulent political moment witnessed in Brazil in recent years may have influenced the results, since other factors could have been considered by participants at the time to choose one of the candidates available.
In addition, there are the natural limitations of online questionnaires, among them: impossibility of providing assistance to participants or unfavorable conditions at the time to fill them up. The questionnaire responding rate was low, although it was within the expected average; it was lower than any other questionnaire application method (Evans and Mathur, 2018). Other natural limitations linked to research conducted on the Internet, such as hard time including incentives for response sending and low data reliability, even when control questions are included, must be highlighted (Evans and Mathur, 2005; Litvin and Kar, 2001). Furthermore, the present study did not include specific variables related to economic voting or clientelism, which could contribute to explaining voters’ decision-making.
Given the aforementioned limitations, it is recommend running similar studies with larger samples. Yet, the questionnaire shall be applied not only on the Internet, but also in person, and it would perhaps help to achieve higher response accuracy. Furthermore, factor ‘religion’ could be added to the study, because it would trigger the ever-recurring discussion about the religion/politics entanglement. Another limitation of our study is that we analyzed the degree of utilitarianism and deontology without considering the underlying factors that explain the respondents’ ethical orientation [see discussion in Gawronski and Beer (2017)]. Future studies could include these factors, incorporating them as endogenous variables in systems of equations. Finally, given the ongoing polarization wave between the left and the right wing in Brazil, future studies could make in-depth analysis about whether the causal link between these groups have different impact on municipal, state and national elections.
Data availability statement
The dataset and codes used in this study can be found in Mendeley Data: https://dx.doi.org/10.17632/56t7wxhmtt.1.
Ethics statement
This research was exempt from submission to the research ethics committee, according to Resolution CNS No. 510/2016, which addresses quantitative survey-type investigations. All participants signed the electronic consent form to participate in the research.
Author contributions
JS: Conceptualization, Formal analysis, Investigation, Methodology, Writing – original draft, Writing – review & editing. LR: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing. CG: Formal analysis, Methodology, Writing – original draft, Writing – review & editing.
Funding
The author(s) declare that financial support was received for the research, authorship, and/or publication of this article. National Council for Scientific and Technological Development (CNPq Grant No. 313387/2021–4).
Conflict of interest
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
The author(s) declared that they were an editorial board member of Frontiers, at the time of submission. This had no impact on the peer review process and the final decision.
Generative AI statement
The authors declare that no Generative AI was used in the creation of this manuscript.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
Footnotes
1. ^https://agenciadenoticias.ibge.gov.br/en/agencia-news/2184-news-agency/news/39816-in-2023-wage-bill-and-per-capita-household-earnings-hit-record
References
Agerberg, M. (2020). The lesser evil? Corruption voting and the importance of clean alternatives. Comp. Pol. Stud. 53, 253–287. doi: 10.1177/0010414019852697
Alesina, A., and Passarelli, F. (2019). Loss aversion in politics. Am. J. Polit. Sci. 63, 936–947. doi: 10.1111/ajps.12440
Arvate, P., and Mittlaender, S. (2017). Condemning corruption while condoning inefficiency: an experimental investigation into voting behavior. Public Choice 172, 399–419. doi: 10.1007/s11127-017-0452-x
Atzmüller, C., and Steiner, P. M. (2010). Experimental vignette studies in survey research. Methodology 6, 128–138. doi: 10.1027/1614-2241/a000014
Barros, L., Goldszmidt, R., and Pereira, C. (2019). Why do voters choose corrupt candidates? The role of ideology on cognitive mechanisms. Int. J. Public Opinion Res. 32, 676–692. doi: 10.1093/ijpor/edz041
Barros, D. F., Sauerbronn, J. F. R., and Ayrosa, E. A. T. (2012). Representações do eleitor: revendo teorias e propondo novos caminhos. Revista de Administração Pública 46, 477–491. doi: 10.1590/S0034-76122012000200007
Bélanger, É., and Meguid, B. M. (2008). Issue salience, issue ownership, and issue-based vote choice. Elect. Stud. 27, 477–491. doi: 10.1016/j.electstud.2008.01.001
Berinsky, A. J., Margolis, M. F., and Sances, M. W. (2014). Separating the shirkers from the workers? Making sure respondents pay attention on self-administered surveys. Am. J. Polit. Sci. 58, 739–753. doi: 10.1111/ajps.12081
Boas, T. C., Hidalgo, F. D., and Melo, M. A. (2019). Norms versus action: why voters fail to sanction malfeasance in Brazil. Am. J. Polit. Sci. 63, 385–400. doi: 10.1111/ajps.12413
Bonifácio, R. (2013). A afeição dos cidadãos pelos políticos mal-afamados: identificando os perfis associados à aceitação do rouba, mas faz no Brasil. Opinião Pública 19, 320–345. doi: 10.1590/S0104-62762013000200004
Bøttkjær, L., and Justesen, M. K. (2021). Why do voters support corrupt politicians? Experimental evidence from South Africa. J. Polit. 83, 788–793. doi: 10.1086/710146
Brady, F. N., and Wheeler, G. E. (1996). An empirical study of ethical predispositions. J. Bus. Ethics 15, 927–940. doi: 10.1007/BF00705573
Brei, Z. A. (1996). Corrupção: dificuldades para definição e para um consenso. Revista de Administração Pública 30, 64–77.
Breitenstein, S. (2019). Choosing the crook: a conjoint experiment on voting for corrupt politicians. Res. Politics 6, 1–9. doi: 10.1177/2053168019832230
Carreirão, Y. S. (2008). Opiniões políticas e sentimentos partidários dos eleitores brasileiros. Opinião Pública 14, 319–351. doi: 10.1590/S0104-62762008000200003
Carreras, M., and Vera, S. (2018). Do corrupt politicians mobilize or demobilize voters? A vignette experiment in Colombia. Latin Amer. Polit. Soc. 60, 77–95. doi: 10.1017/lap.2018.25
Colombo, C., and Steenbergen, M. R. (2021). “Heuristics and biases in political decision making” in The Oxford encyclopedia of political decision making. ed. D. P. Redlawsk (Oxford: Oxford University Press).
Conway, P., and Gawronski, B. (2013). Deontological and utilitarian inclinations in moral decision making: a process dissociation approach. J. Pers. Soc. Psychol. 104, 216–235. doi: 10.1037/a0031021
Costa, C. F. (2002). Razões para o utilitarismo: uma avaliação comparativa de pontos de vista éticos. ethic@-An International Journal for Moral. Philosophy 1, 155–174.
De Vries, C. E., and Solaz, H. (2017). The electoral consequences of corruption. Annu. Rev. Polit. Sci. 20, 391–408. doi: 10.1146/annurev-polisci-052715-111917
Dias, J., and Bento, F. (2011). Corrupção e teoria econômica. Revista Economia and Tecnologia 7, 55–62. doi: 10.5380/ret.v7i3.26615
Evans, J. R., and Mathur, A. (2005). The value of online surveys. Internet Res. 15, 195–219. doi: 10.1108/10662240510590360
Evans, J. R., and Mathur, A. (2018). The value of online surveys: a look back and a look ahead. Internet Res. 28, 854–887. doi: 10.1108/IntR-03-2018-0089
Fligstein, N., and Dauter, L. (2007). The sociology of markets. Annu. Rev. Sociol. 33, 105–128. doi: 10.1146/annurev.soc.33.040406.131736
Gaines, B. J., Kuklinski, J. H., and Quirk, P. J. (2007). The logic of the survey experiment reexamined. Polit. Anal. 15, 1–20. doi: 10.1093/pan/mpl008
Gawronski, B., and Beer, J. S. (2017). What makes moral dilemma judgments “utilitarian” or “deontological”? Soc. Neurosci. 12, 626–632. doi: 10.1080/17470919.2016.1248787
Gonçalves, C. P., and Ayrosa, E. A. T. (2022). Em que medida o autoritarismo e o discurso político violento influenciam a intenção de voto e o engajamento em campanha de candidatos de direita? Um experimento de vinhetas. Revista Eletrônica de Ciência Administrativa 21, 406–429. doi: 10.21529/RECADM.2022015
Hainmueller, J., Hangartner, D., and Yamamoto, T. (2015). Validating vignette and conjoint survey experiments against real-world behavior. Proc. Natl. Acad. Sci. 112, 2395–2400. doi: 10.1073/pnas.1416587112
Huber, J., Payne, J. W., and Puto, C. (1982). Adding asymmetrically dominated alternatives: violations of regularity and the similarity hypothesis. J. Consum. Res. 9, 90–98. doi: 10.1086/208899
Jenke, L., and Huettel, S. A. (2020). Voter preferences reflect a competition between policy and identity. Front. Psychol. 11:566020. doi: 10.3389/fpsyg.2020.566020
Kahneman, D., Knetsch, J. L., and Thaler, R. H. (1991). Anomalies: the endowment effect, loss aversion, and status quo bias. J. Econ. Perspect. 5, 193–206. doi: 10.1257/jep.5.1.193
Kumar, K., and Strandholm, K. (2002). American business education effect on the ethical orientation of foreign students. J. Educ. Bus. 77, 345–350. doi: 10.1080/08832320209599687
Lago, I. (2019). A research agenda in elections and voting behavior in a global and changing world. Front. Polit. Sci. 1, 1–13. doi: 10.3389/fpos.2019.00001
Litvin, S. W., and Kar, G. H. (2001). E-surveying for tourism research: legitimate tool or a researcher’s fantasy? J. Travel Res. 39, 308–314. doi: 10.1177/004728750103900309
Love, E., Salinas, T. C., and Rotman, J. D. (2020). The ethical standards of judgment questionnaire: development and validation of independent measures of formalism and consequentialism. J. Bus. Ethics 161, 115–132. doi: 10.1007/s10551-018-3937-8
Megías, A., de Sousa, L., and Jiménez-Sánchez, F. (2023). Deontological and consequentialist ethics and attitudes towards corruption: a survey data analysis. Soc. Indic. Res. 170, 507–541. doi: 10.1007/s11205-023-03199-2
Middleton, K. L., and Byus, K. (2004). The relationship between pre-charismatic leader emergence and ethical attitudes and behaviors: Hope for the future? J. Appl. Manag. Entrep. 9, 86–100.
Miranda, L. F. (2018). Unificando os conceitos de corrupção: uma abordagem através da nova metodologia dos conceitos. Revista Brasileira de Ciência Política 25, 237–272. doi: 10.1590/0103-335220182507
Perannagari, K. T., and Chakrabarti, S. (2020). Analysis of the literature on political marketing using a bibliometric approach. J. Public Aff. 20:e2019. doi: 10.1002/pa.2019
Philipson, T. J., and Snyder, J. M. Jr. (1996). Equilibrium and welfare in markets with induced preferences. J. Polit. Econ. 104, 147–155.
Piurko, Y., Schwartz, S. H., and Davidov, E. (2011). Basic personal values and the meaning of left‐right political orientations in 20 countries. Political Psychology 32, 537–561. doi: 10.1111/j.1467-9221.2011.00828.x
Resende, M. M., Porto, J. B., Gracia, F. J., and Tomás, I. (2024). Unethical behavior at work: the effects of ethical culture and implicit and explicit moral identity. Ethics & Behavior 34, 438–457. doi: 10.1080/10508422.2023.2243632
Rose, J. (2018). The meaning of corruption: Testing the coherence and adequacy of corruption definitions. Public Integrity 20, 220–233. doi: 10.1080/10999922.2017.1397999
Scammell, M. (1999). Political marketing: Lessons for political science. Political studies 47, 718–739. doi: 10.1111/1467-9248.00228
Shleifer, A., and Vishny, R. W. (1993). Corruption. The Quarterly Journal of Economics 108, 599–617. doi: 10.2307/2118402
Simon, H. A. (1955). A behavioral model of rational choice. The quarterly journal of economics 69, 99–118. doi: 10.2307/1884852
Simon, H. A. (1956). Rational choice and the structure of the environment. Psychological Review 63, 129–138. doi: 10.1037/h0042769
Simonson, I., and Tversky, A. (1992). Choice in context: Tradeoff contrast and extremeness aversion. Journal of Marketing Research 29, 281–295. doi: 10.1177/002224379202900301
Smith, A., and Pulver, S. (2009). Ethics-based environmentalism in practice: Religious-environmental organizations in the United States. Worldviews: Global Religions, Culture, and Ecology 13, 145–179. doi: 10.1163/156853509X438580
Smith, K. B., Alford, J. R., Hibbing, J. R., Martin, N. G., and Hatemi, P. K. (2017). Intuitive ethics and political orientations: Testing moral foundations as a theory of political ideology. American Journal of Political Science 61, 424–437. doi: 10.1111/ajps.12255
Stoetzer, L. F., and Zittlau, S. (2020). Vote choice and the nonseparability of economic and social issues. Public Opinion Quarterly 84, 158–170. doi: 10.1093/poq/nfaa005
Tversky, A. (1994). Contingent preferences: Loss aversion and tradeoff contrast in decision making. Japanese Psychological Research 36, 3–9. doi: 10.4992/psycholres1954.36.3
Tversky, A., and Simonson, I. (1993). Context-dependent preferences. Management Science 39, 1179–1189. doi: 10.1287/mnsc.39.10.1179
Vera, S. B. (2020). Accepting or resisting? Citizen responses to corruption across varying levels of competence and corruption prevalence. Political Studies 68, 653–670. doi: 10.1177/0032321719868210
Wang, A. H. E., and Chen, F. Y. (2019). Extreme candidates as the beneficent spoiler? Range effect in the plurality voting system. Political Research Quarterly 72, 278–292. doi: 10.1177/1065912918781049
Weitz-Shapiro, R., and Winters, M. S. (2017). Can citizens discern? Information credibility, political sophistication, and the punishment of corruption in Brazil. The Journal of Politics 79, 60–74. doi: 10.1086/687287
Whelan, J., Goode, M. R., Cotte, J., and Thomson, M. (2016). Consumer regulation strategies: Attenuating the effect of consumer references in a voting context. Psychology & Marketing 33, 899–916. doi: 10.1002/mar.20927
Keywords: mayors, aversion to extremes, ethics, elections, corruption, utilitarianism, deontology
Citation: Silva JOS, Rossoni L and Gonçalves CP (2025) Ethical orientation and choice context of mayoral candidates accused of corruption in Brazil: randomized experiment based on vignettes. Front. Polit. Sci. 6:1505042. doi: 10.3389/fpos.2024.1505042
Edited by:
Nicolas Tsapatsoulis, Cyprus University of Technology, CyprusReviewed by:
Anthony Costello, Liverpool Hope University, United KingdomRodrigo Horochovski, Federal University of Paraná, Brazil
Copyright © 2025 Silva, Rossoni and Gonçalves. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
*Correspondence: Luciano Rossoni, bHJvc3NvbmlAdW5iLmJy