Elections are the main mechanism for politicians to account to citizens for their actions. In turn, members of the public show a firm rejection of corruption and frequently consider it to be one of Spain’s main problems (Pring, 2017). Therefore, and following a “reward and punishment” logic, we would expect that politicians who abuse power and get involved in corruption scandals would receive less support at the polls. Is this the case? 

To address this question, this experimental study is based on an online survey conducted in 2020 among a representative sample of 2,540 Spaniards, from whom, initially, information on party preferences was obtained. The experiment randomly assigned to 567 of the participants the reading of a news item about a corruption scandal involving their preferred party. Another group (control group) with 577 participants read, in its place, a neutral news item on politics or entertainment, without any relationship with corruption or political parties. Next, all those surveyed answered a question regarding their likelihood of voting for their favourite party. 

Figure 1 reflects how, effectively, the likelihood of voting for their favourite party is lower among those exposed to news about corruption. Among those surveyed, the likelihood of voting for their favourite party is, approximately, some 5.5% lower than among those assigned to the control group, who have not read that same news.

This first analysis suggests that, once corruption scandals involving certain parties are made public, the parties should receive less electoral support. However, this evidence is contradicted by the fact that, frequently, corruption is not punished at the ballot box (De Vries and Solaz, 2017). 

This apparent contradiction may be due to the fact that, in the real world, not all citizens are exposed to news about the corruption scandals affecting their favourite party. In fact, in the results presented in figure 1, as well as in the majority of prior experimental studies, there was random assignation to some participants of the reading of said information, without taking into account the fact that, in reality, many of them would avoid reading it. This ignores the fact that one of the pre-requisites for the electoral punishment of corruption is that individuals receive news regarding it. If citizens are not informed about the irregularities committed by their representatives, we cannot expect these to have an impact on their vote.

What happens when citizens can choose the news that they consume? It could be expected that, under these circumstances, not all citizens would opt to read about the corruption cases affecting their favourite party. 

The attitudes and political beliefs of citizens are fundamental when selecting news. Thus, some individuals tend to expose themselves to news that questions or contradicts their political preferences. Others opt for completely avoiding political news and only consume entertainment contents. Moreover, it is probable that individuals who do want to read about corruption will have different socio-political characteristics to those opting to read other types of news. Numerous studies indicate that individuals with greater political knowledge, greater interest in politics and higher partisan identification consume more political news but, in turn, are less likely to modify their attitudes and behaviours based on new information that they receive (Zaller, 1992). All of this has significant implications for the electoral punishment of corruption. However, the proportion of individuals who expose themselves to news about corruption and how these are differentiated from other citizens are two aspects that have not been taken into account in previous studies.

To obtain information on these questions, we have introduced here two innovative elements with regard to previous seminal studies on the effects of corruption (Anduiza et al., 2013; Winters and Weitz-Shapiro, 2013). Firstly, at the start of the survey and knowing the favourite political party of those surveyed, they were shown different headlines and asked which news item they would prefer to read, with these being the options: a news item about a corruption case affecting their favourite party, a news item on the good performance of their favourite party, a neutral political news item (making no reference to any political party) and two news items on entertainment.  

Next, 819 people of the 2,540 surveyed were randomly assigned to read the news selected by themselves in accordance with their preferences. This scenario, in which each individual selects for him or herself the news that he or she wishes to read, is more realistic than that in which all survey respondents are randomly exposed to news that perhaps they would not read in the real world.

Figure 2 shows how only 15.5% of those surveyed choose to read about a corruption case involving their favourite party. The majority (some 48.2%) prefer to read about the good performance of their preferred party. Furthermore, some 33.2% of those surveyed avoid reading about politics and opt, instead, for news about entertainment. Finally, only 3.1% choose to inform themselves about neutral political issues, which do not make reference to any political party. The fact that only a minority of citizens decide to expose themselves to information about the corruption affecting their favourite party may represent a serious limitation for the electoral punishment of corruption. 

In addition, as revealed by figure 3, the profile of those surveyed who prefer to read each of these types of news items is different. Those who read about corruption or the good performance of their favourite party are politically more sophisticated: they have greater interest in and knowledge about politics. These survey respondents also identify more intensively with their favourite party, especially those who prefer to read about the party’s good performance. In contrast, those who opt to read about neutral political issues, or issues related with entertainment, identify in a much weaker way with their favourite party. The latter clearly constitute the group with the least political interest and knowledge. As for the ideological profile of each group, people who prefer to read about corruption are slightly more left-wing in their views, whereas those who opt for neutral political issues are more right-wing.

The differences in the profiles of individuals who select each one of these types of news items can have consequences for the electoral punishment of corruption. This is so because prior studies have demonstrated that individuals with greater political sophistication and a more defined partisan identification are more reluctant to modify their political attitudes and opinions (Zaller, 1992). Knowledge and prior preferences lead them to select and process new information, in such a way that they can avoid cognitive dissonances. Therefore, even when they are exposed to news about corruption cases that involve their favourite political party, it is possible that individuals with greater political interest and knowledge, and strongly identified with the same, are less likely to modify their opinions regarding this party. In contrast, individuals who are less sophisticated and weakly identified with their favourite party may be more receptive to this news and modify their opinions to a greater extent.

Figure 4 connects the types of news survey respondents would like to read with the likelihood of voting for their favourite party. The orange bars correspond to the group assigned to read a news item about corruption, independently of their news preferences. Thus, the orange column situated to the right of the graph shows the likelihood of voting for their favourite party among those surveyed with a preference for reading news about entertainment but that ultimately, were randomly assigned to read about corruption in their favourite party. The blue bars correspond to the group which was allowed to choose the news that they wanted to read and that, therefore, could read a news item in line with their preferences. For example, in this group, those with a preference for reading news items on the good performance of their party could effectively read such a news item. 

In other words, the blue bars estimate what happens between individuals with different news preferences when, as in real life, they can select the news to which they are exposed, while the orange bars represent what would happen in a parallel world in which we could ensure that these same individuals were exposed to news about the corruption of their favourite party. Therefore, the differences between each pair of orange and blue bars enable us to estimate how electoral punishment would change if people with different news preferences were exposed to news about this issue.

Among those who prefer to read the news about corruption in their favourite party, the likelihood of voting for the same party is very similar in the two groups. This is to be expected since both those assigned randomly to reading the news about corruption (orange bar) and those who chose this news in accordance with their preferences (blue bar) were exposed to the same information. 

In contrast, the lack of differences between the two groups who wish to read about the good performance of that party is, a priori, surprising. The likelihood of voting for their favourite party is very similar among those who wanted to read news about the good performance of their political party, but were exposed to news about corruption and those who, in effect, could opt to read about the good performance of their party. This could be due to the high level of interest and political knowledge and intensity of party identification of these citizens, which makes them more reluctant to modify their opinions regarding their favourite party, even if they receive negative information about it. This suggests that, in the case of these individuals, the selection of news does not limit the punishing of corruption, since, even if they received news regarding corruption in their preferred party, they would not modify their support for it to any great extent. 

The situation is clearly different among those with a preference for informing themselves on neutral political issues or who simply opt for entertainment news. In both cases, the survey respondents who, in accordance with their preferences, read neutral political news or news about entertainment, showed themselves to be substantially more likely to vote for their favourite party than those who shared the same news preferences, but were exposed to a news item on corruption of the party. It is worth highlighting that the result is more robust in those with a clear preference for entertainment. Among these, the study demonstrates clearly that their exposure to news about corruption would result in greater electoral sanctioning of this behaviour.

The results indicate that the fact that citizens selectively expose themselves to political news and avoid news related with the corruption of their favourite party limits the electoral punishment of this behaviour. This would be explained, especially, by the presence of citizens who do not expose themselves to this information and opt for entertainment news instead of politics. Among these citizens we would observe a greater increase in the punishing of corruption if they received news regarding this serious problem. Taking as a reference the proportion of survey respondents who opt for entertainment news (results presented in figure 2) and the change in the voting intention observed among these when exposed to news regarding the corruption of their preferred party (figure 4), this study estimates that if these citizens were exposed to this type of news, the support for said party in the elections would be reduced by some 4.6%. 

The results also lead to a pessimistic conclusion. Among the people with greater political interest and knowledge who do opt to read about politics (although to a large extent about the good performance of their favourite party), the margin for increasing accountability is limited. Among these individuals, exposing themselves to a news item about a corruption scandal that affects their favourite party will barely affect their likelihood of voting for the same, which suggests that their attitudes and behaviour are less malleable. 

Therefore, information campaigns that may be designed around increasing accountability and electoral punishment of corruption will be more effective if directed at people who normally would not expose themselves to political news and who opt, instead, for entertainment contents.

ANDUIZA, E., A. GALLEGO and J. MUÑOZ (2013): «Turning a blind eye: experimental evidence of partisan bias in attitudes toward corruption», Comparative Political Studies, 46. 

DE VRIES, C.E., and H. SOLAZ (2017): «The electoral consequences of corruption», Annual Review of Political Science, 20.

PRING, C. 2017. People and Corruption: Europe and Central Asia Global Corruption Barometer. Berlin: Transparency International.

WINTERS, M.S., and R. WEITZ-SHAPIRO (2013): «Lacking information or condoning corruption: when do voters support corrupt politicians?», Comparative Politics, 45.

ZALLER, J. (2014): La naturaleza y  los orígenes de la opinión pública, Madrid: Centro de Investigaciones Sociológicas (ed. or. 1992)