Media Credibility and Voter Penalization of Corrupt Politicians in Latin America

Abstract

There has been a significant growth of social media as a means to inform oneself about politics. This article explores the consequences of this trend on the credibility audiences attribute to news exposing corrupt politicians and on their willingness to penalize the exposed politicians in elections. The study focuses on ten Latin American cities and employs a randomized control trial using experimental data embedded in a survey. Through this method, credibility and penalization levels are compared between state communications, newspapers, named journalists on social media, and anonymous journalists on social media. The article’s key findings demonstrate that corruption reports published on social media are deemed less credible than those published by state auditors and newspapers. This effect is exacerbated when the source of the report is anonymous. In addition, reports on corruption published on social media by anonymous sources have a negative effect on voter penalization of corrupt politicians.

1. Introduction

Grand corruption in Latin America remains high, particularly in the public sector. Latin Americans are aware of the harm corruption entails, but the region often fails to penalize corrupt politicians during elections (Berniell et al. 2019). The determinants of voter attitudes are naturally complex, multidimensional, and context specific. For instance, Lula (da Silva) came out of the 2005 vote-buying Menselão scandal somewhat unscathed and won the 2006 presidential election in Brazil with more than 60% of the votes (in the second round). On the other hand, Keiko Fujimori’s corruption charges (based on money laundering) a month prior to the 2021 Peruvian presidential election arguably cost her the presidency.

One of the key mechanisms to increase the electoral penalization of corruption is media and information availability (Ferraz and Finan 2008; Chong et al. 2012; Bobonis 2016; Arias et al. 2019). Nonetheless, the majority of studies concerned with information availability and the penalization of corruption assume that the information provided in the media is deemed credible by audiences. However, many citizens in Latin America do not have much trust in their political institutions which often becomes reflected in lower perceptions of credibility in traditional media (such as newspapers and government communications). ¹ At the same time, journalists who wish to report on sensitive issues may resort to alternative publishing channels such as social media, an outlet which is often associated with less perceived credibility and oversensationalization of news (Viviani and Pasi 2017; Karlsen and Aalberg 2021; Lin et al. 2016) but has also become the second most popular source of information about politics in Latin America (CAF 2018). Our article focuses hence on the idea that the electoral penalization of corruption aided by media exposure not only depends on the availability of information, but also on the credibility that audiences attribute to that information.

The vast majority of studies that focus on news media and their perceived credibility are not within the context of political corruption. Nonetheless, there are few notable exceptions. One of the most relevant studies linking information credibility and voter penalization of corruption is an article concerned with the ability that citizens have to discern credible sources and how this may affect electoral accountability in Brazil (Winters and Weitz-Shapiro 2017). It is one of the few studies acknowledging that information availability may not be sufficient to ensure the penalization of corrupt politicians in the ballots; the credibility audiences attribute to the available information matters as well. Winters and Weitz-Shapiro conclude that politically sophisticated citizens are better at discerning credible sources and are thus better equipped to act against political malfeasance or corruption. Hence, their study proposes that increasing education (jointly with information availability) is an effective mechanism to increase the electoral penalization of corruption. Our research builds on this earlier analysis. However, while the focus of Winters and Weitz-Shapiro (2017) lies in the determinants of individual ability to better discern credibility, our intent is to assess how variation in the source of information (particularly that of social media) influences perceived credibility and the penalization of corruption. Two other relevant analyses are the ones by Muñoz et al. (2016) and Botero et al. (2015) on how credibility of corruption reporting is affected by political party affiliations in Spain and the type of messenger in Colombia, respectively. These studies demonstrate the increased relevance of information credibility in the fight against corruption through media and elections.

Conceptual framework on media trust and hypotheses. Our research design and hypotheses are based on the conceptual framework of the causes of media trust by Fawzi et al. (2021), which itself relies on a broad review of the relevant (theoretical and empirical) literature. ² They classify correlates of trust in news media into media-related, social, and political categories.

Media-related correlates. Fawzi et al. (2021) claim that “people… develop more differentiated attitudes, trusting some outlets while distrusting others” (Fawzi et al. 2021, 162). A number of factors simultaneously drive differences in trust across different news media. Naturally, perceptions of accuracy play a pivotal role in trust (Prochazka 2020; Bachmann and Valenzuela 2023) and social media are often accused of being prone to deficiencies in fact checking (Rembischevski and Caldas 2023). Perceptions on the expertise of journalists also matter (Schielicke et al. 2014) and, therefore, anonymous social media are likely to suffer even more from reduced credibility. The presentation style is another influencing factor, with more sensational and emotionally charged styles (common in social media) linked to reduced credibility (Molyneux and Coddington 2019; Goyanes et al. 2021) but not always (see García-Perdomo et al. 2024). On the other hand, perceptions of bias are closely associated with distrust towards media (Fico et al. 2004; Serrano-Puche et al. 2020; Mont’Alverne et al. 2021), which can affect especially state communications and newspapers (as these are often perceived as biased in favor of the government or certain political parties, Hughes and Lawson 2004). Knowledge of media ownership can further enhance perceptions of bias (Ashley et al. 2010); this can hence potentially increase trust in social media outlets without a clear ownership status vis-à-vis newspapers and government-owned outlets (especially in weaker democracies, Tsfati and Ariely 2014; Rodríguez and Zechmeister 2018).

Social correlates. Under the framework of Fawzi et al. (2021), individual social characteristics influence media trust. For instance, people with a higher educational attainment may be more critical towards media credibility and better suited to assess journalistic integrity (De Freitas et al. 2021); however, evidence on the relationship between education and media trust is mixed (e.g. the studies by Gronke and Cook 2007 and Tsfati and Ariely 2014 suggest a positive and negative link respectively). In addition, how one relates to others also seems to matter; individuals who largely feel that they “do not fit in with the rest” or tend to distrust others will also exhibit limited media trust (and especially so when information is not easily verifiable, see Tsfati and Ariely 2014).

Political correlates. Fawzi et al. (2021) discuss the role of individual political characteristics. A higher level of political sophistication (familiarity with the broader political spectrum/scene) can be linked to a better ability to discern credibility; this often translates into an overall higher trust in media but simultaneous distrust towards anonymous social media outlets (that are then considered untrustworthy, Hansen and Kim 2011; Tsfati and Ariely 2014). ³ In parallel, political cynicism is often closely linked to media cynicism (Pinkleton et al. 2012), the extent of which likely depends on the verifiability of media information.

Consequences of media trust. Fawzi et al. (2021) rely on their framework to predict consequences of trust in media. High media trust is typically associated with the formulation of stronger affections and reactionary responses (in line with the messages depicted in the corresponding news outlet, see Ladd 2012; Salzman 2012); this could in principle strengthen the intention to penalize corrupt politicians when exposed in the news. On the other hand, individuals with a weaker trust in media may be more inclined to vote more uncritically and in line with their party affiliations (Ladd 2012) which can mitigate any reactionary response to media exposures of political corruption.

Our study makes use of elements of this framework on the causes of media trust to evaluate factors influencing the credibility attributed to corruption reports in Latin America and the intention of electoral penalization (of exposed corrupt politicians). The article hence devises three specific research questions, each of them tailored to address specific aspects of the changing media landscape in Latin America and their possible implications on credibility and the penalization of corrupt politicians. These questions and their respective hypotheses are laid out as follows:

1. 1. Is social media (versus other, more traditional, media) considered less credible when exposing incidences of political corruption in Latin America? This question builds on the “media-related correlates” of Fawzi’s framework that claims that media trust is likely to differ across media outlets. Here we explore which media outlets and sources are considered most credible for Latin Americans in the context of journalists exposing political grand corruption. The media choices include state communications, newspapers, journalists publishing on social media, and anonymous journalists publishing on social media. Particular attention is paid to social media, given their rise in popularity. This first research question leads to two hypotheses:

   H1a. Social media is expected to be perceived as less credible than other media outlets (often due to their denationalization of information and less fact checking). This hypothesis assumes that the public is aware of this characteristic of social media and would hence deem them less credible.

   H1b. Anonymity on social media should hinder credibility even further, given that such reports miss a key credibility cue: the author’s name. This hypothesis assumes that respondents can discern credibility cues and will attribute less credibility to anonymized journalistic reports.

2. 2. Does the intention of electoral penalization (of exposed corrupt politicians in Latin America) depend on the medium of communication? This question builds on the “consequences of media trust” of Fawzi’s framework, which can again be media specific. Here we explore which media outlets and sources lead to a higher willingness of penalizing corruption among the Latin American public. Again, the article pays special attention to social media in this comparison, given that media credibility is likely to be an important mediating factor.

   H2. The article hypothesis that citizens are less likely to penalize corrupt politicians when corruption is exposed on social media than when it is exposed through other media channels. This statement is based on the expectation that social media is perceived as less credible, as stated in H1a, and the assumption that higher credibility attributed to a corruption report would drive a higher penalization of corruption.

3. 3. Are there any individual-specific factors that increase media credibility and the electoral penalization of corruption? This question builds on the “social and political correlates” of Fawzi’s framework that claims that media trust is mediated by individual social and political characteristics. The article tests a set of individual determinants to find whether they have a significant effect on media credibility and/or voter penalization. This test is carried out for every medium and source relevant to the experiment.

   H3. The individual-specific determinants that increase media trust are one’s educational attainment, level of political sophistication (interest in politics) and perception that concerns about corruption are shared with other fellow citizens. Given this effect on credibility, the respondent’s willingness to penalize the exposed politician is also expected to increase accordingly.

Our article makes a twofold contribution to the field. First, our analysis has a much broader geographical focus compared to earlier studies. Our article covers ten major Latin American cities, being the first comprehensive comparative study on media sources, credibility, and voter penalization for Latin America. Second, studies in media credibility and corruption often focus on traditional media channels. Our article considers traditional, state owned, and social media sources. This is particularly relevant given the rapid decline in newspaper readership and the parallel rise in social media use, combined with the perception that social media often disseminate fake unverified news (impacting the credibility of legitimate reporting on corruption). In order to make these two contributions, the article conducts a randomized control trial using an experiment embedded in a survey titled Encuesta Corporación Andina de Fomento (ECAF). The survey, which focuses on corruption and public office integrity, was conducted in 2018 by Corporación Andina de Fomento (CAF), a development bank based in Latin America.

The results demonstrate that information disseminated through social media can be perceived as less credible, validating the article’s first hypothesis. Despite this, and in response to the second research question, this does not appear to significantly influence voter penalization (with the exception of anonymized journalistic reports). Regarding the third research question, results support earlier findings in the literature confirming that education and minimal political sophistication have a positive effect on credibility. However, when assessing the willingness to penalize corrupt politicians, education loses its significance. Given these results, the article can make informed anti-corruption policy recommendations.

The article is structured as follows. Section 2 provides the context and literature relevant to the article’s focus. Section 3 describes the data used in the analysis. Section 4 explains the methodology and specification relevant to each research question. Section 5 consists of a discussion of the results and Section 6 concludes.

2. Literature Review

2.1 Corruption in Latin America

Our article focuses on grand corruption in the public sector, a form of corrupt behavior that typically involves fewer people but larger exchanges of money (Prasad et al. 2019). Focusing on grand corruption allows us to assess the determinants of electoral penalization, given that it is more pertinent to high-level politics, as opposed to bureaucratic or petty corruption. We narrow the conceptualization of corruption by focusing on the misappropriation of funds, using positions of power in political spheres for private gain. It is important to make this distinction given that petty corruption does not tend to be a major focal point for media reports (given its broader systematic and cultural nature). On the other hand, media-exposed cases of public sector corruption in the form of funds misappropriation typically relate to the behavior of specific culprits (politicians in our case).

Public sector corruption in Latin America consistently ranks as one of the highest in the world. Except for Chile and Uruguay, all Latin American countries rate poorly in corruption indices. Transparency International rates the Latin American region with a Corruption Perception Index (CPI) of 35 on average compared to an average global CPI of 43 points, meaning that Latin America has a higher corruption level than the world average (Transparency International 2021). This is corroborated by the V-Dem’s index on corruption in the public sector (V-Dem 2021). Corruption at such high levels entails significant developmental costs. It negatively affects economic growth, investment, and government expenditure (Mauro 1995). This has been corroborated by a large number of studies (Shaw et al. 2011; d’Agostino et al. 2016; de Paulo et al. 2022; Everhart et al. 2009; Sharma and Mitra 2019). Additionally, corruption affects many other developmental factors, as in the case of mortality rates (Escaleras and Register 2016), environmental conservation (Pellegrini 2011), health (Anbarci et al. 2009) and education (Heyneman et al. 2008; Duerrenberger and Warning 2018).

Data from a 2018 survey by CAF shows that Latin Americans are aware of the harm corruption entails for economic progress, social equality, and the quality of public services. They even single out corruption as the main issue of concern their countries face. In light of this, some turn to electoral penalization as a viable anti-corruption solution. However, Latin Americans have in large part demonstrated a striking lack of penalization towards corrupt politicians in elections (Berniell et al. 2019). Scholars have attempted to understand this phenomenon. In some cases, voters expect the benefits of certain politicians being in office to exceed the costs of their corrupt acts (Winters and Weitz-Shapiro 2013). In other instances, voters living in countries with weak and corrupt institutions feel resignation towards political institutions and do not think their vote will change the state of corruption in the public sector (Agerberg 2019). Most scholars, however, seem to argue that Latin Americans fail to penalize corrupt politicians due to lack of information in the media. The effect of the media (or information availability) on voter penalization has been confirmed by several seminal studies. Ferraz and Finan (2008) demonstrate that the availability of information from state audits increases electoral penalization of corrupt politicians in Brazil, while the propagation of these audits in the media increases this effect. Bobonis et al. (2016) follow this line by showing that Puerto Rican municipalities with timely audits before elections have considerably lower levels of corruption. The dissemination of audit results is paramount for these effects to materialize. Additionally, a study on Mexico specifies that the availability of information about corrupt incumbents does increase electoral accountability, but it decreases voter turnout and does not increase the support for challengers’ parties (Chong et al. 2012). In addition, recent studies find a negative relationship between corruption and social media penetration, even in contexts of restricted press freedom (Enikolopov et al. 2018; Jha and Sarangi 2017). The higher the social media use is, the lower the metrics of corruption. Furthermore, highly connected networks enable voters to coordinate among themselves and share information to penalize corrupt politicians (Arias et al. 2019). Hence, the connectivity and openness that social media offers may indicate a newfound opportunity to curb corruption.

2.2 Media Credibility

Credibility refers to the expectation that someone’s word can be relied on. This expectation is built on that person’s reputation, which depends on consistently providing truthful information (Sobel 1985, 557). The study of credibility in media can be categorized into two areas (see Golan 2010, 10): studies on medium credibility (i.e., media channels such as television, newspapers, etc.) and studies on source credibility (i.e., the messenger such as a speaker, organization, author, etc.). In the case of Latin America, social media is the second most popular channel of information about politics (CAF 2018). Tsfati (2010, 38) claims that the rising tendency of alternative media outlets is driven by heightened skepticism towards traditional media. However, non-mainstream media often have defining characteristics that tend to clash with journalistic professionalism or traditional news values (Tsfati 2010, 26). It is no secret that information posted online often does not entail an exhaustive factual verification (Flanagin and Metzger 2000, 516). Given this, several studies focus on social media credibility topics such as fact checking, evaluating indicators of credibility, curbing the spread of misinformation, and the recent phenomenon of fake news (Viviani and Pasi 2017; Yaqub et al. 2020). On the contrary, in the case of traditional media, Asak and Molale (2020) demonstrate that newspapers have low instances of fake news and generally abide by professional journalism principles. The literature presents a scholarly debate on whether social media is deemed more (or less) credible than traditional sources such as newspapers or television broadcasts. On the one hand, Karlsen and Aalberg (2021) propose that news published on social media are deemed less credible by audiences, particularly when politicians are intermediary messengers. On the other hand, Johnson and Kaye (2010) claim that audiences consider the internet more trustworthy than traditional media, particularly during electoral and campaigning seasons. In addition to the study of medium credibility, scholars argue that the source of information also plays a key role in credibility. Miller and Kurpius (2010) assessed the determinants of credibility based on the characteristics of messengers in television broadcasts; Meyer et al. (2010) suggest that it is the author’s perceived expertise that influences the perceived credibility by news audiences. In the case of social media, Tandoc (2019) claims that the credibility attributed to news depends on the messenger (an organization vs. friends), particularly when the motivation to engage with the news is high. If motivation is low, audiences tend to engage in more heuristic cues that affect their ability to discern credibility in sources. Similarly, other studies assessed how heuristics related to the source and the recency of the post affect the perception of credibility on social media (Lin et al. 2016; Westerman et al. 2014).

3. Data

The article uses data and an experiment embedded in a household survey titled Encuesta Corporación Andina de Fomento (ECAF). This survey was conducted in 2018 by CAF and focuses on corruption and public office integrity. ECAF covers a sample of 9,621 individuals spread evenly across ten major Latin American cities. These include Buenos Aires (Argentina), La Paz (Bolivia), São Paulo (Brazil), Bogotá (Colombia), Quito (Ecuador), Mexico City (Mexico), Lima (Peru), Panama City (Panama), Montevideo (Uruguay), and Caracas (Venezuela). Respondents were given a monetary incentive for their time and effort spent in answering the household survey. The sample is randomized by geographically delimited strata in order to ensure the spatial spread of the sample. The strata are defined by neighborhoods in every city of interest. Within each stratum, the number of sample points are defined in proportion to the population of that stratum. These sample points are randomly selected. The sample is balanced for observable variables, such as gender, age, and country of origin.

The data in this article is thus categorized in two sections: (i) an experiment embedded in ECAF 2018 and (ii) survey data from ECAF 2018. Both sections are explained in detail below.

3.1 The Experiment

One particular experiment from ECAF 2018 motivates this study. This experiment can be found in question 43 in the ECAF 2018 questionnaire. ⁴ In this experiment the respondent is told about a report exposing irregularities in the purchase of materials authorized by a government minister. This report is published in different media channels/sources, which are considered treatments in the experiment. The respondents are thus randomized across three treatment groups and one control group. The control group is told that a state auditor releases the report. The first treatment group is told that the nation’s most popular newspaper releases the report. The second treatment group is told that a named journalist publishes the report on social media (Facebook, Twitter, Instagram, etc.). The third treatment group is told that a group of anonymous journalists publish the report on social media. The researcher then asks two questions:

Q1: On a scale from 1 to 5, how credible would you find this report?

R: 1 – not credible; 2 – not very credible; 3 – somewhat credible; 4 – quite credible; 5 – very credible

Q2: On a scale from 1 to 5, how likely is it that you would vote for this minister if he/she puts himself/herself forward as a candidate in the next elections?

R: 1 – not likely; 2 – not very likely; 3 – somewhat likely; 4 – quite likely; 5 – very likely

The experimental data comprise 7,558 observations for the credibility question and 7,561 observations for the penalization question. ⁵ It should be noted that the experiment is subject to some limitations. One of the limitations arises from the fact that this is an experiment embedded in a survey. This may raise concerns about the survey questions prior to the experiment, given that they may condition the responses to the treatments. As the survey proceeds, the respondent’s awareness of corruption may become more salient and thus increase one’s inclination to believe a report about a corrupt politician regardless of the media source. In other words, one’s perception of credibility may be overestimated. Similarly, one’s willingness to penalize a corrupt politician might increase throughout the survey. However, while this may affect the overall mean of the outcome variables, it should not affect the differences between treatments. Another limitation is that the treatments in the experiment are not made salient in any way to the respondents. The report and government minister that are central to the questions in the experiment are both hypothetical and immaterial. While this ensures that no political or media figure is attached to the conceptualization of the experiment, this does leave the concept of the corruption report open to interpretation. The article hence makes the assumption that the corruption report is conceptualized similarly by respondents (in other words, that the understanding of the treatment is homogenous across all respondents). Additionally, the article assumes that respondents do not make associations between the hypothetical government minister and relevant political figures in their respective countries.

The article hence utilizes three treatment groups and one control group. The control group includes the state as a source of information to assess how much people trust their governments compared to other communication outlets (i.e., to contrast credibility in journalistic vs. non-journalistic government information). The first and second treatment groups include traditional and social media channels (in both cases with information provided by named journalists) to address the implications of the rise of social media, which is relevant to the first and second hypotheses of the article. Finally, the third treatment involves anonymous journalists on social media, which is crucial for two reasons. First, the inclusion of this treatment is relevant because it can indicate the level of attention audiences dedicate to credibility cues (see Section 2.2). There are two possible opposing forces here. On the one hand, anonymous social media postings may be associated with reduced credibility as this prevents identifying the identity of the sender and verifying the accuracy of information. On the other hand, anonymous postings may be considered more credible since this allows journalists to express their opinions more freely and without the risk of external interferences or retribution. Second, this third treatment is likely to be particularly relevant in environments (not atypical in many Latin American contexts) where freedom of expression is insufficiently protected and when journalists who investigate and report on corruption cases become harassed and threatened (Pennachio 2021; Nalvarte 2020; Aruth Sturm 2016); this hence incentivizes some journalists to publish their investigations anonymously. As a matter of fact, several Latin American journalists make use of anonymous blogging platforms (such as telegra.ph or write.as) or post in more common social media platforms (e.g. Facebook) with the use of pseudonyms.

In order to facilitate the interpretation of results and the analysis in general, the variables for credibility and penalization were both redefined into binary variables. Instead of carrying out the study on a Likert scale, the two outcome variables of interest are stipulated in binary terms. The definition of these two variables can be seen in Table 1 above.

Table 1. Binary Variables for Credibility and Penalization

Source: Authors’ analysis using data from ECAF 2018 (CAF 2018).

3.2 ECAF 2018 Data

The analysis of the experiment is complemented with additional variables from ECAF 2018. The purpose of this data is to analyze certain individual characteristics that could influence the respondents’ perception of credibility and their willingness to penalize corruption within each treatment group. Here we focus on three key determinants (variables).

The first determinant is the maximum level of education attained by the respondent. The survey’s original variable includes 13 levels of education. For the purpose of our analysis, this is simplified into a binary variable distinguishing between those who did not complete secondary school versus those who attained an educational level of secondary school and above. This distinction does not only serve as a simplification for the analysis, but it is also pertinent to the context of Latin America. According to World Bank Data (2020), the average rate of people who completed secondary school in the countries relevant to this study is 49 percent. Hence, attaining a secondary school qualification marks a balanced distinction between those who are better educated from those who are not so well educated. As Weitz-Shapiro and Winters (2017) demonstrated, education plays an important role in people’s ability to discern credible sources. It is also linked to their level of political sophistication.

The second determinant of interest relates to a minimal level of political sophistication. This is measured in the survey by asking if the respondent is familiar with concepts of right-wing and left-wing politics. The addition of this variable allows us to test Winters and Weitz-Shapiro’s theory of political sophistication and its effect on credibility. It should be noted that there are some missing values for the political sophistication variable (which, however, only amount to 2 percent of the total observations and hence pose no major concern).

The third determinant involves the respondent’s perception of his or her fellow citizens’ concern about corruption. In other words, the variable responds to the question, “Do you believe your fellow citizens consider corruption events when casting their vote?” (CAF 2019). This variable aims to indicate the perception respondents have about the public’s general attitude towards corruption in their country (and its possible penalization by the average voter at the ballot box).

Table 2 presents the descriptive statistics for all variables of our analysis (The Supplementary Material Section presents the corresponding subgroup city values).

Table 2. Descriptive Statistics

Source: Authors’ analysis using data from ECAF 2018 (CAF 2018).

4. Specification and Methodology

The central methodology in this article, given the experimental nature of the data, is a randomized control trial (RCT). As mentioned in Section 3, ECAF 2018 was carried out following a systematic randomization method. This ensures the minimization of biases in the assignment of the treatments and allows the study to infer causal differences between groups. The experiment design allows the article to assess the effects of different media/sources of information on two outcomes: (i) perception of credibility of a corruption report, (ii) the willingness to penalize a corrupt politician. In all cases, these outcomes are assessed at a regional level (i.e., Latin America) with country fixed effects. This allows the article to assess regional trends but also to control for each country’s particular characteristics, given that the context of corruption and freedom of expression is country specific. Also, in all cases, Uruguay acts as a reference category given its exemplary position with low corruption levels and high freedom of expression in relation to all other sample countries.

4.1 Average Treatment Effects

We first obtain the Average Treatment Effect (ATE) per outcome to evaluate the differences in treatment groups and the control group (regarding perceived credibility and penalization intent). Since the sample and the treatment assignation are randomized, we assume that the Average Treatment Effect on the Treated (ATT) is equal to the ATE. In other words, the selection and heterogeneity effects in the model should not be a concern. Also, spillover effects from treatments are not a concern since the treatments are assigned instantaneously in the household survey. There is no time gap between the assignment and the response to the treatment for any spillover effects to materialize.

There are two models, one for each outcome variable:

$${\rm{Credibilit}}{{\rm{y}}_{{\rm{ij}}}} = {\rm \alpha} + {\rm \beta _1}{{\rm{T}}_1} + {\rm \beta _2}{{\rm{T}}_2} + {\rm \beta _3}{{\rm{T}}_3} + {\rm \delta _{ij}} + {\rm \varepsilon _{{\rm{ij}}}}$$

$${\rm{Penalization}_{{\rm{ij}}}} = {\rm \rho} + {\rm \gamma _1}{{\rm{T}}_1} + {\rm\gamma _2}{{\rm{T}}_2} + {{\rm \gamma} _3}{{\rm{T}}_3} + {\rm \eta _{{\rm{ij}}}} + {{\rm\xi} _{{\rm{ij}}}}$$

where i and j refer to individuals and countries respectively. The coefficients of interest in these models are those pertaining to the treatment dummy variables (T₁, T₂, T₃). Each of these are compared to the control group, which acts as a reference category. The coefficients thus provide the treatment effects for every treatment group. The models also include constants α/ρ, country fixed effects δ/η, and error terms ϵ/ξ.

4.2 Heterogeneous Treatment Effects

Our analysis defines specific heterogeneous groups to assess whether they incur an additional effect on credibility or penalization based on their treatment. In this case, the article assesses whether there is a heterogeneity effect in the ATE tested above. For this purpose, we extend the aforementioned specifications by including interaction terms between each treatment group and a vector of person-specific variables X_ij capturing the relevant characteristic per heterogeneous group (T_kX_ij), as well as the X_ij variables alone. The coefficient of interest in this case is the one pertaining to the interaction term, which provides the additional effect of the group variable on the treatment effect. This specification allows us to assess whether certain characteristics of respondents enhance or diminish treatment effects on credibility and/or penalization.

5. Results

5.1 Average Treatment Effects: Credibility and Penalization

Treatment effects per media outlet are presented in Table 3; we observe that newspapers are perceived to be the most credible source of information for corruption reports. Newspapers have an increased probability (of 3 percentage points) of being deemed more credible than state communications (this result is statistically significant at the 10 percent level). Newspapers may be deemed more credible given that they represent a medium that is independent from the state. This means that newspapers are free to publish corruption reports without being influenced by the state, while state communications may choose to conceal information about a corruption case (especially if a member of their own political party is involved). Despite this reasoning, the difference in credibility is small, which may indicate that traditional media and state communications are similar or intertwined in the mind of the public.

Table 3. ATE on Credibility and Penalization

Treatment effects on credibility and penalization per treatment group compared to control group (state auditor). p-values in parentheses: *p < 0.10, **p < 0.05, ***p < 0.01. Country fixed effects in all regressions. Source: Authors’ analysis using data from ECAF 2018 (CAF 2018).

In the case of the second treatment, a journalist publishing on social media faces a reduced probability of being deemed credible compared to state communications by 3.9 percent. This is statistically significant at the 5 percent level. This suggests that audiences trust state communications more than journalists on social media. This finding corroborates the article’s first hypothesis. There are two possible reasons behind this effect. First, as demonstrated by the literature, social media is a medium that is prone to deficiencies in fact checking, source verification, and other credibility indicators. This result may thus indicate that the public is aware of these negative characteristics. Second, information about politics is often sensationalized on social media. Stories that induce emotions, especially anger, are promoted in social media platforms to drive engagement (Ghosh 2021). Hence, this result may indicate that respondents are aware of social media’s tendency towards sensational news and attribute correspondingly less credibility to them. Nonetheless, it should be noted again that the difference in credibility is rather small in size. This may be because state communications also face their own limitations in terms of credibility, namely the incentive to conceal information as discussed above.

When we compare the first and second treatments, we observe that social media is regarded as a less credible medium. This comparison allows the article to weigh in on the debate regarding whether social media or newspapers are deemed the most credible (and reveals that, in the context of corruption reports in Latin America, traditional media earn increased credibility vis-à-vis social media). This confirms the article’s first hypothesis. Nonetheless, these effects are rather small to allow taking a definite stance in the ongoing debate (i.e., audiences do not seem to make stark credibility distinctions between media channels).

The credibility of state communications (and newspapers) is higher compared to the credibility of social media. This implies that state action against corruption is more effective in terms of establishing credible information. Hence, despite the power of social media as an influential channel of information dispersion, state communications and newspapers continue to play a major role in exposing corruption from a credibility standpoint. A number of factors simultaneously drive differences in trust across different news media. Naturally, perceptions of accuracy play a pivotal role in trust (Prochazka 2020; Bachmann and Valenzuela 2023) and social media are often accused of being prone to deficiencies in fact checking (Rembischevski and Caldas 2023). Perceptions on the expertise of journalists also matter (Schielicke et al. 2014) and, therefore, anonymous social media are likely to suffer even more from reduced credibility. The presentation style is another influencing factor, with more sensational and emotionally charged styles (common in social media) linked to reduced credibility (Molyneux and Coddington 2019; Goyanes et al. 2021) but not always (see García-Perdomo et al. 2024). On the other hand, perceptions of bias are closely associated with distrust towards media (Fico et al. 2004; Serrano-Puche et al. 2020; Mont’Alverne et al. 2021).

Finally, the results offer a comparison between Treatments 2 and 3 to evaluate the effects of anonymity. An anonymous journalist on social media has a reduced probability of 4.4 percent of being deemed credible compared to state communications (or a 7.7 percent difference in credibility when compared with newspaper announcements). This result is statistically significant at the 1 percent level. If we compare this to the negative effect of 3.9 percent mentioned earlier for Treatment 2, it is clear that source anonymity does affect credibility (albeit marginally). These results corroborate the article’s first hypothesis regarding the effect of anonymity on perceived credibility. The small difference in effects between the two treatments suggests that anonymity does not seem to provoke a strong negative response from audiences. This may indicate that, to some extent, audiences in countries where freedom of expression is low are more prone to trust anonymous journalistic pieces. This assumes that the public is aware of the risks journalists face when investigating and publishing about corruption (Aruth Sturm 2016; Nalvarte 2020; Pennachio 2021). In these instances, anonymity would not hinder credibility since it is regarded as a necessity. Another possible reason behind audiences accepting anonymity to a certain extent is the lack of education and ability to discern credibility cues in journalistic pieces.

What is interesting about these findings is that social media continues to be the second most popular media outlet, after television, in Latin America for people to inform themselves about politics (CAF 2018). However, at the same time, our findings demonstrate that social media is considered a less credible source of information compared to other media outlets that are less popular in use. This raises the question on the importance the public places in perceived credibility when selecting information sources for politics (in relation to other criteria, such as the convenience and ease of access that social media typically offer).

In the case of treatment effects for electoral penalization (Table 3, last column), the coefficients for newspapers and social media are small and statistically insignificant. This result indicates that variation in media does not matter in driving the electoral penalization of corruption, which implies that the rise of social media does not necessarily limit demand for penalization. The third treatment of anonymity on social media is associated with a (modestly) reduced probability of penalizing the corrupt politician by 3.5 percent, compared to the control group. This result is statistically significant at the 5 percent level. One of the possible reasons behind this is that anonymity presents the greatest lack of credibility given the results demonstrated above. As respondents do not deem the report credible given its source, the information about the corrupt politician would not warrant a penalizing action.

The results for Treatment 3 nod at the idea that credibility and penalization have a positive correlation, as demonstrated in Figure 1. This scatterplot illustrates the coefficients for each treatment group compared to the control group for both credibility and penalization. A positive relationship between the two variables is clear across all treatment groups. However, it is worth noting again that differences in the penalization rate across media types are relatively modest (which is possibly associated with the also relatively modest differences in the corresponding credibility rates).

Figure 1. ATE on Credibility and Penalization.

Treatment effects on credibility and penalization per treatment group compared to control group. Treatment 1: top right-hand corner; Treatment 2: top left-hand corner; Treatment 3: bottom left-hand corner. Source: Authors’ analysis using data from ECAF 2018 (CAF 2018).

5.2 Heterogeneous Treatment Effects: Credibility and Penalization

Motivated by earlier literature findings, we conduct HTE tests for three key determinants: i.e., the level of education, political sophistication, and the concern of fellow citizens about corruption. Each group yields at least one statistically significant effect on credibility or penalization, except for the influence of fellow citizens. Results are presented in Table 4.

Table 4. Heterogenous Treatment Effects

Heterogeneous treatment effects on credibility and penalization per treatment and control groups. Coefficients belong to interaction terms between treatment groups and group variable. p-values in parentheses: *p < 0.10, **p < 0.05, ***p < 0.01. Source: Authors’ analysis using data from ECAF 2018 (CAF 2018).

In the case of education, those respondents who have not finished secondary school have a 5 percent lower probability of deeming a report credible when published on social media. This is compared to those who completed secondary school (or higher) and is statistically significant at the 10 percent level. Those who are better educated are probably in a better position to evaluate the journalistic integrity of the source and discern credibility, even in the case of social media (i.e., by searching for cues in the text or verifiable sources rather than base their perception of credibility on the medium alone).

In addition, being minimally politically sophisticated decreases the probability of finding the report credible by 6.7 percent among those who were told that the report was published by an anonymous journalist on social media. This result is statistically significant at the 5 percent level. This seems intuitive, given that politically sophisticated people are better able to discern credibility and would distrust anonymous publications. This corroborates Winters and Weitz-Shapiro’s findings (2017).

In the case of penalization, only political sophistication has statistically significant effects. For those who were told that the report was published in a newspaper, being politically sophisticated increases the probability of penalizing the politician by 6.2 percent. This is statistically significant at the 5 percent level. Moreover, if the report is published by an anonymous journalist on social media, being politically sophisticated decreases the probability of penalization by 4.3 percent. This is statistically significant at the 10 percent level. Similar to the results on credibility, Treatment 3 has a negative effect on the penalization of corruption. It could be argued that given the lack of substantiation in the report’s credibility, politically sophisticated audiences are reluctant to penalize the politician in question.

6. Conclusion

Our article successfully assesses three hypotheses concerned with media, credibility, and the electoral penalization of corruption. The first hypothesis concerned the credibility of social media; overall, social media seems to have a negative effect on credibility, although the effect is rather small (this effect is exacerbated when the author of the report is anonymous). The article’s second hypothesis, which is concerned with the effect of social media on the penalization of corruption, is only partially supported. Only the third treatment, anonymity on social media, has a significant negative effect on penalization. This result indicates that penalization is not affected by the medium, but rather by the source of reports. As for the third hypothesis, HTE results support findings in the literature concerned with education and minimal political sophistication.

Given these results, we put forward some key policy recommendations. The first recommendation is to promote initiatives that aim at improving social media as an informational space. The continuous use of social media is inevitable, despite often being deemed a less credible medium compared to others. Several studies have addressed opportunities for improvement, including fact checking, increasing journalistic integrity and source verification, among others (Viviani and Pasi 2017; Yaqub et al. 2020). Some social media platforms have promised to carry out such improvements, but their goals fall short of addressing the scale of the problem (Ghosh 2021). This matter is subject to a heated debate, questioning whether improving the quality and credibility of information should be left to the volition of social media companies (Ghosh 2021). The second recommendation echoes the findings of Winters and Weitz-Shapiro’s article (2017). It is evident that the role of political sophistication is key in driving the penalization of corruption. Increasing education and political sophistication enables a population to advance their skills in discerning credible sources and assess the journalistic integrity of corruption reports. This is worrying, since approximately 49 percent of young Latin Americans do not complete their secondary school requirements (World Bank Data 2020). Reducing social inequality and poverty, as well as enhancing appropriate education-oriented policies, are at the forefront of improving educational attainments.

We have several extensions of our current analysis in mind. First, the article does not consider other communication channels outside the written press and social media, such as radio or television. This is a missed opportunity given that television is the most popular media channel to inform oneself about politics in Latin America (CAF 2018). This omission is due to the construction of the ECAF survey and its exclusive focus on written communication (and omission hence of other communication channels, as in the case of television and radio). Second, future research could broaden the scope of our analysis by assessing social media’s persuasive characteristics beyond credibility, since these have a stronger impact on emotions than traditional media. Sensationalized news and emotion-inducing posts are more prone to drive action in audiences (as in the case of protesting, voting, etc.). This presents an opportunity to assess these elements and rethink how we interact with information and how this affects our actions in political settings. In addition, future research could extend our study by factoring in the effect of political polarization. Societies that are highly polarized in politics often experience a rise of alternative or fringe media that caters to such polarization. As Strömbäck et al. (2020) mention, fringe media resulting from demand for more partisan news outlets undermines trust in media. In light of the contemporary rise in political polarization, this would be worth exploring.

Last, there is an important concern of external validity (and the ability to generalize our findings and apply these to a broader context, given the decontextualized nature of the survey experiment). Our study does not treat politics in depth and is not concerned about the affiliation respondents have to certain political parties (and how this affects the credibility they attribute to media channels or their willingness to penalize corruption). It is important to note that media channels and authors also have political affiliations and, in many cases, a loyal following among readers. Similarly, respondents are presented with hypothetical information about an unnamed government minister making reference to an incidence of political corruption. In reality, voters trade off accusations of corruption against partisan interests and assessments of the candidate’s competence and prior performance (i.e., their leniency towards political corruption is likely to be dependent on the politician’s characteristics and other anticipated costs and benefits). In order to gain external validity, the study would have to shift its focus to one particular context and truly reflect the political intricacies and affiliations attached to the electoral process, the media, and the penalization of corruption. ⁶ Despite this caveat, our analysis makes a compelling contribution to the field by providing an encompassing assessment of Latin America as a region and the role of social media in the connection between media credibility and the electoral penalization of corruption.