In the Face of Disinformation: To Publish or Not to Publish in the Vaza Jato Case

Abstract

This article analyses journalistic decisions in the face of disinformation, focusing on the case of Vaza Jato in Brazil. Drawing on a mixed-methods approach—combining critical discourse analysis of online articles with semi-structured interviews with two editors—the study explores how two ideologically contrasting newspapers (Folha de S.Paulo and Gazeta do Povo) framed and justified their editorial positions regarding the publication of hacked content. The findings reveal distinct narrative strategies, degrees of epistemological openness, and levels of institutional trust in the judiciary and political actors. The results also show how editorial decisions are shaped by broader concerns about professional legitimacy, audience trust, and the ambiguous boundary between journalism and disinformation. This article contributes to research on disinformation, editorial ethics, and media trust, proposing an analytical framework applicable to other high-risk communication contexts.

1. Introduction

Democracy relies on independent and trustworthy media to inform citizens with truthful and accurate content (e.g., Knight Foundation, 2023; Michailidou et al., 2023). However, trust in media dropped between 2021 and 2023 in most countries (Watson, 2024b), confirming a global tendency that had already intensified after the onset of the pandemic (Serrano-Puche et al., 2023). In the United States of America (USA), this erosion of trust is particularly acute, driven not only by pandemic-related dynamics but also by the political polarization that escalated after 2016, when populist narratives of media discrediting gained traction (Fawzi, 2018; Knight Foundation, 2023). Half of Americans now believe that news organizations deliberately mislead them (Knight Foundation, 2023). The problem of disinformation gets worse because, in addition to leading to judgments far from the truth that motivate decisions with real effects, it exerts an influence that persists in people’s reasoning beyond its correction—the continued influence effect (Ecker et al., 2022).

However, it appears that Brazil is bucking this trend, given that, at the end of 2022, around 46% of internet users surveyed in Brazil expressed their trust in the media as an institution that does what is right (an increase of 5 percentage points compared to 2018, when the previous president Jair Bolsonaro was elected) (Guttmann, 2023). These data align with the findings of the OECD Trust Survey, which show that the news media were the third most trusted institution in Brazil (OECD iLibrary, 2024). However, a 2023 study conducted in four countries found a decline in trust in Brazil, where just 39.8% say they trust news “somewhat” or “completely” (Banerjee et al., 2023, online).

This article focuses on the journalistic approach taken by two Brazilian portals to the Vaza Jato reports first published by The Intercept Brasil (TIB) in June 2019. Vaza Jato focuses on the publication of leaked conversations between Sergio Moro, then a federal judge, and prosecutors from the Federal Public Prosecutor’s Office (MPF) who are members of the Lava Jato task force, which began in 2014 and was considered the largest police operation against corruption in Brazil (Gabardo et al., 2021). The conversations held between 2015 and 2018 suggested questionable practices and sparked debates about the impartiality of justice. In 2019, Sergio Moro became Minister of Justice in Jair Bolsonaro’s government, raising new questions about his performance, since it was Moro who convicted Luiz Inacio Lula da Silva, preventing him from running in the 2018 elections against Jair Bolsonaro.

Vaza Jato secured a distinctive place in Brazilian journalism and reshaped media coverage of Lava Jato, prompting wide legal-political repercussions and industry self-scrutiny (Gabardo et al., 2021; Oliveira, 2020), as well as unprecedented cross-newsroom collaboration with TIB (Duarte, 2020). We focus on this case because it crystallizes (i) a high-stakes test of newsroom verification under conditions of asymmetric access to leaked material (Ireton & Posetti, 2018; Wardle & Derakhshan, 2017); (ii) competing editorial logics in Brazil’s media field, as a national legacy outlet with primary access (Folha de S.Paulo) and a regional digital-first outlet without access (Gazeta do Povo) responded to the same disclosures in a polarized environment (Fawzi, 2018; Hiltunen, 2022); and (iii) the boundary work between journalism and disinformation, where provenance, context and power are as consequential as factual accuracy (Tandoc et al., 2018; Zeng & Brennen, 2023). Rather than surveying disinformation content, we examine editorial decision-making—when and why editors choose to publish (or refrain from publishing) contested, illegally obtained material—because gatekeeping choices determine whether ambiguous information is amplified or contained (Marwick & Lewis, 2017; Tandoc et al., 2018). To avoid over-breadth, we streamline our initial motivations into two contextual justifications: (a) documented links between newsroom constraints (economic, professional, political) and the circulation of misleading or decontextualized claims (Allen et al., 2020; Amazeen, 2020; Brennen, 2020; Cook et al., 2014; Hendrickx & Van Remoortere, 2023; Robertson et al., 2023); and (b) the democratic stakes of trust in news, especially when verification and public-interest arguments must be weighed under legal/ethical uncertainty (Banerjee et al., 2023; Jakobsson & Stiernstedt, 2023; Knight Foundation, 2023; McQuail, 2010; OECD iLibrary, 2024).

The framing of this study within disinformation research is not only pertinent but necessary. The Vaza Jato case lies within a complex informational territory—a liminal zone where verified data, digital piracy, and strategic manipulation intersect. This type of empirical object exemplifies the emerging configurations of disinformation, which extend beyond the deliberate spread of falsehoods to encompass opaque domains, where source provenance, dissemination context, and embedded interests are as relevant as factual accuracy (Wardle & Derakhshan, 2017).

The case of Vaza Jato invites not only a discussion on disinformation, but also a deeper epistemological reflection on the role of journalism in interpreting and legitimizing ambiguous information. Drawing on Ricoeur’s (1970) concept of the hermeneutics of suspicion, this study acknowledges that media narratives often operate within a framework of distrust, where the provenance, intention, and power embedded in information are not taken at face value. This approach is particularly relevant when journalistic actors must assess whether or not to publish material obtained through questionable means. Such editorial decisions require a critical negotiation between public interest and ideological framing—reinforcing the idea that disinformation is not merely about falsity, but about contextual trust and interpretative authority (Ecker et al., 2022; Jakobsson & Stiernstedt, 2023).

The uniqueness of this work lies in its focus not on disinformation content per se, but on the editorial decision to publish ambiguous or controversial material. This dimension—underexplored in the literature—is critical, as it is through journalistic gatekeeping that potentially dubious information is either amplified or contained. As Tandoc et al. (2018) argue, media are not merely victims of disinformation; they are also co-constructors of its circulation and legitimacy. By focusing on the intersection of ethics, verification, and ideology, this article proposes a meaningful theoretical contribution.

Furthermore, by analyzing the editorial strategies of two ideologically distinct newspapers, this study shows that disinformation is not a binary phenomenon, but a field of symbolic struggle in which public trust, epistemic authority, and the democratic role of journalism are at stake. In this sense, it aligns with scholarship on the ecology of disinformation (Marwick & Lewis, 2017), by demonstrating that the core risk lies not only in the falsity of information, but in the political management of doubt.

The aim of this article is to offer a contribution at the level of “operational theory”, as this refers to the practical ideas gathered and applied by media professionals in conducting their own media work, serving to guide solutions to fundamental tasks (McQuail, 2010), including how to select news, how to deal with the value of information and potential disinformation.

To compose a literature review on the thematic focus in approach, a Boolean search was carried out in July 2025 on the EBSCOhost database, “the leading provider of research database” (https://www.ebsco.com/), using the keywords “disinformation or fake news or misinformation” AND “trust or confidence” AND “newspaper or newspapers or news media” AND “Vaza Jato or corruption”. The search generated no useful results, which shows the pioneering nature of this thematic proposal.

In theoretical terms, this study is based on the theory of agenda-setting, which focuses on the selection and prioritization of certain topics, and the theory of information gatekeeping, which focuses on the decisions of news producers, as well as studies on disinformation and trust in the media. Specifically, this study aims to 1. assess the narratives of the Folha de S.Paulo and Gazeta do Povo portals on Vaza Jato that focus on the truth or lack thereof; 2. understand how the discourse of the news and editorials of the Folha de S.Paulo and Gazeta do Povo portals on Vaza Jato addresses the value of the information published; and 3. assess the position of the editors of the Folha de S.Paulo and Gazeta do Povo portals that motivated their decisions in relation to the ethical commitments of disseminating information on Vaza Jato. In line with these objectives, we formulated three research questions that directly mirror them:

RQ1: How do Folha de S.Paulo and Gazeta do Povo construct narratives about Vaza Jato in terms of truth versus falsehood?

RQ2: In their news and editorials, how do both outlets assess the value of the published information?

RQ3: What editorial positions expressed by the editors-in-chief motivated their decisions to publish (or not) information on Vaza Jato, in light of their ethical commitments?

To fulfill each of the objectives and answer the research questions, narrative analysis, discourse analysis and interview are used respectively. The method is mixed and the type is explanatory sequential that prioritizes the qualitative dimension, with interviews being conducted with the purpose of also explaining the results of the techniques previously applied.

2. Between Disinformation and Trust in the Media—The View of Journalism

Among journalists, academic communities and the public, there is growing interest and concern about the proliferation of disinformation. This phenomenon becomes a problem, especially in political events, involving electoral decisions, and in health crises. Although there is no absolute consensus on certain aspects of the concept of disinformation (e.g., informativeness and intentionality), one conceptual aspect that has been widely accepted is its falsity (e.g., Zeng & Brennen, 2023).

On the other hand, trust in the media is a central issue in academic research, especially in journalism studies. Trust includes emotional and cognitive dimensions and functions as a deep assumption that sustains social order, being “functionally necessary for the continuance of harmonious social relationships” (Lewis & Weigert, 1985, p. 969).

Public trust is sensitive to the transformation of the communication environment in recent decades and to social, economic, cultural and technological changes (Serrano-Puche et al., 2023). The study by Serrano-Puche et al. (2023), based on the qualitative method (focus group), shows that one of the main reasons for Spanish citizens’ distrust of the media is the perception that the news is biased for political or economic reasons. The COVID-19 pandemic, in which disinformation abounded, influenced attitudes towards the media and the way news is consumed. The authors note that after the pandemic, trust in the media decreased and, as a result, alternative sources of information were sought.

In a media ecosystem characterized by the hybridization between the logics of classic and digital media (Hiltunen, 2022) and by the high information supply that makes users’ information consumption choices difficult, there is an abundance of partisan and alternative sources that disseminate content attacking traditional media, accusing them of not being trustworthy. Ideologically, populist parties and movements have adopted a discourse of discrediting journalism, which has contributed to increasing skepticism and animosity towards the media. For example, the results of a representative German survey (Fawzi, 2018) predominantly confirmed a relationship between a populist worldview and negative attitudes towards the media, which are perceived as a distanced elite that neglects the interests of citizens. Also based on 31 semi-structured interviews conducted with Finnish journalists, Hiltunen (2022) finds that the societal and political polarization were identified by the interviewees as both symptoms and driving forces of the erosion of public trust in journalism, going hand in hand with increasingly contested and fragmented public sphere.

Journalism researchers have found a relationship between the media and the rise of disinformation: 1. Traditional journalism’s circulation and advertising revenues have been declining (e.g., Lipka & Shearer, 2023), and this decline has been exacerbated by the rise of digital media and news consumers’ unwillingness to pay for it (e.g., Kormelink, 2022). As a result, revenues have fallen, which has been reflected in layoffs and closures (e.g., Grynbaum et al., 2024). To reduce costs, journalists, instead of producing content from their own research, are rewriting content from news dispatches and media releases (Brennen, 2020). These changes mean that less independent, high-quality news is produced (Hendrickx & Van Remoortere, 2023), and that uneven access to sources or leaks can generate a news vacuum, which in turn creates conditions for conjecture, partisan framing, and the circulation of misleading or decontextualized claims (Brennen, 2020; Ireton & Posetti, 2018; Marwick & Lewis, 2017; Wardle & Derakhshan, 2017; Zeng & Brennen, 2023). When analyzing hard news content, often focused on information essential for citizens to make informed electoral decisions, the weakening of quality news media and their contribution to news pluralism is a particularly worrying trend that Hendrickx and Van Remoortere (2023) propose should be the subject of academic and regulatory scrutiny. 2. Potential emotional reactions from news stories can encourage or facilitate misinformation (Horner et al., 2021). The economics of emotion theory proposes that fake news headlines are created to evoke emotional responses in readers that will make them interact with the article in a way that allows the creator to make a profit. The study by Horner et al. (2021) showed that U.S. participants were more likely to believe headlines aligned with their beliefs (e.g., liberals are more likely to believe negative news about conservatives) and reacted with more negative emotions to headlines attacking their party, being more likely to suppress such news. Participants’ emotional reactivity was associated with behavioral intentions. Thus, participants who reported high levels of emotion were more likely to take actions that would spread or suppress fake news, while participants who reported low levels of emotion were more likely to ignore or tune out fake news. 3. Certain journalistic practices can unintentionally facilitate misinformation. In this study, objectivity is not treated as a simplistic “two-sides” formula. Following the literature, objectivity is a professional norm anchored in verification, independence, methodological transparency, and fairness in representation—an historically contingent ideal that structures newsroom routines rather than a mere balance of opposing statements (Ireton & Posetti, 2018; Schudson, 2001; Waisbord, 2013). When enacted narrowly as both-sides-ism, it can grant artificial equivalence to unsubstantiated claims and thereby legitimize problematic information. In parallel, the adoption of outrage as a news style—by design eliciting strong affect—can further enable the spread of misleading content; and negativity in headlines is associated with higher consumption: for an average-length headline, each additional negative word increases click-through by 2.3% (Robertson et al., 2023). 4. The media can promote misinformation through opinion pieces (Cook et al., 2014), sponsored content (Amazeen, 2020) or uncritical repetition of politicians’ statements. 5. Some academics go further and hypothesize that “the origins of public misinformedness [are] more likely to lie in the content of ordinary news or in the avoidance of news altogether as they are in overt fake news” (Allen et al., 2020).

Given the serious effects of the problem, the literature has been dedicated to identifying strategies for building audience trust in the news media. For example, the Reuters Institute’s Trust in News Project (Banerjee et al., 2023) examined cases in the North (USA and United Kingdom [UK]) and South (Brazil and India) using a mixed method. The majority of participants from the publics want news coverage to be “fair, accurate and impartial”, but since perceptions of these characteristics are generally subjective, adopting effective solutions that consider coverage with these attributes is complex (Banerjee et al., 2023). Four approaches to building a relationship of trust with the public are identified (Banerjee et al., 2023): 1. editorial strategies: better aligning news topics with what the audience says they want from trustworthy news outlets; 2. transparency: focusing on communicating ethical standards and newsroom policies, as well as reducing apparent conflicts of interest and biases; 3. management: ensuring journalistic independence and ownership structures that reduce audience skepticism, as well as improving staff diversity; and 4. engagement: proposing initiatives to make the audience feel heard, involving them in news production and responding to their feedback. Audiences are demanding a greater focus of news coverage on ordinary people.

The tendency of the news media to focus on elites in their coverage can contribute to perceptions of the media as disconnected or irrelevant to everyday life. According to news publishers worldwide, the most important initiatives to combat news avoidance and fatigue are better explanation of complex stories (67%), solutions/constructive journalism (44%), more inspiring human stories (43%), diverse presenting reporting team (35%), and simpler language/accessible formats (35%) (Watson, 2024a).

3. Brazil’s Media System and Democratic Context

Brazil’s media ecology has long exhibited high ownership concentration, political instrumentalization and clientelism—features that align more closely with the polarized pluralist pattern in comparative media-systems research—even as Latin American specificities require going beyond the original Western models (Hallin & Mancini, 2004; Waisbord, 2013). Historically, major conglomerates—above all TV Globo—have shaped norms, agendas, and accountability debates within the journalistic field (Porto, 2012). At the same time, scholarship has emphasized structural limits on the press’s watchdog capacity: a social tolerance of corruption, institutional weaknesses, and close links between the state and media ownership, which fostered conflicts of interest and curtailed investigative depth (da Silva, 2000).

Brazil’s democratization since the mid-1980s has been marked by contradictions. On the one hand, liberalization and market expansion created new spaces for pluralism and professional journalism, with national dailies playing visible roles in critical junctures such as the Diretas Já campaign, Collor’s impeachment, and subsequent elections (Matos, 2006). Professionalism, objectivity and watchdog journalism expanded, enhancing credibility, but these gains were consistently constrained by commercialization, sensationalism and denuncismo practices (Matos, 2006). In parallel, broader debates on media democratization underline that Brazil inherited an authoritarian legacy that hinders the consolidation of a genuinely public communication system, with public service broadcasting remaining weak and vulnerable to political instrumentalization compared to the UK or other European democracies (Matos, 2013).

To situate our cases normatively, we incorporate current comparative indicators that classify Brazil as “Free” while documenting vulnerabilities and volatility in civil liberties and institutional quality (Freedom House, 2024). In parallel, recent democracy assessments trace fluctuations in liberal-democracy scores over the last decade, underscoring a constrained electoral democracy rather than an “incipient” one (V-Dem Institute, 2024).

Within this environment, legacy organizations coexist with digital-native investigative actors whose routines, risk tolerance, and editorial missions differ from mainstream newsrooms. This configuration—together with the structural and micro-level constraints highlighted above—is crucial to understand how Brazilian outlets evaluate illicitly obtained leaks, verification burdens, and public-interest claims when faced with ambiguous or politically sensitive material (da Silva, 2000; Global Investigative Journalism Network, 2019; Matos, 2006, 2013; Waisbord, 2013).

4. Disinformation in Brazil and the Vaza Jato Case Study

Brazil is frequently identified as a paradigmatic setting for the circulation of political disinformation, notably around the 2018 presidential election. Empirical studies document large–scale flows of misleading content via messaging apps and social networks—particularly WhatsApp—alongside the role of computational propaganda infrastructures (e.g., social bots) (Arnaudo, 2017; Chaves & Braga, 2019; Resende et al., 2019). These findings help explain how false or decontextualized claims achieved wide reach in a hybrid media system bridging private messaging and open platforms.

A second thread in the Brazilian literature concerns systematic attacks on the professional news media by far-right online activism, which both delegitimizes journalism (“#fakenews”, bias accusations) and strategically appropriates journalistic cues when convenient. This dynamic—well documented in studies of online repertoires—contributes to audience distrust and increases the susceptibility to disinformation frames (Massuchin et al., 2022). Conceptually, these patterns align with the broader “information disorder” framework distinguishing disinformation, misinformation and malinformation and their different logics of harm (Wardle & Derakhshan, 2017).

Within this context, the Vaza Jato disclosures are a critical case. Beginning in June 2019, TIB published leaked Telegram messages attributed to then-judge Sergio Moro and prosecutors from the Lava Jato task force, raising questions about judicial impartiality and triggering wide political-media debate. Contemporary accounts and subsequent analyses reconstruct the newsroom processes behind the series and its repercussions for media–justice relations (Duarte, 2020; Oliveira, 2020; Gabardo et al., 2021).

Editorial responses varied. Folha de S.Paulo, which partnered with TIB, reported with primary access and framed publication around verification, contextual reconstruction and public-interest justifications (see Duarte’s reconstruction and contemporaneous coverage). By contrast, Gazeta do Povo, without primary access, persistently referred to the leaks as “supostas conversas” (“alleged conversations”), foregrounding authenticity disputes and legal issues—an instance of a more defensive gatekeeping stance under uncertainty (Dalmonte & Queiroz, 2020; Gazeta do Povo, 2019a, 2019b). These contrasting strategies illustrate how verification capacity, access asymmetries and editorial identity shape decisions when publication might either counter disinformation (through transparency and corroboration) or inadvertently feed it.

Overall, situating our case study in the Brazilian debate clarifies how structural features (platformized circulation, antagonism toward legacy media) and contingent editorial choices interact in the production- or mitigation—of disinformation effects. The subsection thus provides the national specificity requested by the reviewer while directly framing our comparative analysis of newsroom decision-making.

5. Materials and Methods

A mixed method was used, including narrative analysis (which fulfills the first objective), critical discourse analysis (second objective) and individual, semi-structured, in-depth interviews (third objective).

The corpus to meet the first objective consisted of 828 news pieces about Vaza Jato published in the three months between 9 June and 9 September 2019, on the Folha (530) and Gazeta (298) portals. This period of analysis is justified because it is a full quarter from the publication of the first TIB report on 9 June 2019. It was also the period with the greatest social, political and legal repercussions around the issue. All the categories used in this study were generated concurrently from the literature and the corpus.

Our comparative strategy follows analytical generalization rather than statistical representativeness (Yin, 2014). We analyze two newsrooms selected for maximum contrast in position and routines: Folha (a national legacy title) and Gazeta (a centenarian regional outlet that undertook an early, well-documented digital-first transition) (Estarque, 2017). TIB is not treated as an analytical case: it is the original publisher of the Vaza Jato materials (and, at the time, a content-sharing partner to Folha), providing crucial context for the field but not forming part of our comparative sample (Global Investigative Journalism Network, 2019). This design—legacy national with access/partnership versus regional outlet without primary access—allows us to observe how distinct newsroom ecologies negotiate verification, public-interest justifications, and the risk of instrumentalization when deciding whether (and how) to publish.

5.1. Narrative Analysis

This technique focuses on the difficulties, choices, conflicts and developments of the characters/active agents, in this case those affected by Vaza Jato. Harcourt et al. (2020) explain that, after numerically coding the news texts and identifying the prominent topics, an analysis/reconstruction is carried out in to identify the stories told from this data. This technique therefore takes on a quantitative-qualitative configuration. Based on the narrative policy framework (Shanahan et al., 2018), a narrative analysis was carried out at the meso level of the political actors present in the aforementioned corpus. Using the model of Harcourt et al. (2020) as a reference, narratives that focus on the truth or lack thereof were identified based on impactful events that initiate the story; active agents who respond; and actions taken by these agents as a result of the events. In the reconstruction of the narrative, categories were coded and summarized with descriptive statistical indicators (frequencies and distributions) generated in the Statistical Package for the Social Sciences (SPSS) software; no inferential tests were conducted. These categories characterize corpus-level patterns in time publication (month, numeric day and day of the week of publication), narration mode (framing, statement in first voice, journalistic genre, size of the piece, type of approach, thematic focus, direction, scandal in focus and sources used), protagonists (main active agent) and their actions.

5.2. Critical Discourse Analysis

This technique, which requires depth, was applied to two corpora: a corpus of 45 news pieces (the first stories in each of the scandals covered), 28 from Folha and 17 from Gazeta; and another of six editorials from Folha and six editorials from Gazeta that dealt directly with Vaza Jato and its developments during the quarter under analysis. A three-dimensional dimension was used (

Table 1. Categories.

Categories	Categorization Orientation
Outstanding characteristics of the headline, byline and lead	Highlighted words
Lexical style	Use of words that shape meaning and can reinforce certain opinions
Discursive strategies	E.g., legitimacy: use of sanctions or justifications for the actions of a certain active agent; Positioning: construction of the identity of a person or entity through discourse; Argumentation: use of certain reasoning that leads to a certain opinion.

Source: Authors’ own elaboration based on, e.g., Carvalho (2008).

).

5.3. Interview

Two qualified informants were interviewed (

Table 2. Interviewees.

Media Outlet	Name	Position	Date
Folha	Sergio Davila	Editor-in-chief	25 August 2023
Gazeta	Ewandro Schenkel	Editor-in-chief	19 July 2023

Source: Authors’ own elaboration.

) after they had been contacted to participate and informed about the context and objectives of the research. Both interviews were conducted via video call (Google Meet, version 2023.03.05.515733626.Release), to take into account the availability of the interviewees and to overcome geographical obstacles. The interview guide includes thematic dimensions linked to editorial decisions about publishing the leaked messages, verifying the value of the content and ethical challenges.

The contents of the interviews were transcribed and processed. Framework analysis (Ritchie & Lewis, 2003) was chosen as the basis for the analysis of the data: the stages of familiarization, coding, framework development and interpretation. This framework, using categories, creates a new structure for the data (rather than the full original accounts given by participants) that is helpful to summarize/reduce the data. For ease of reading, quotes have been shortened but still including the essence of interviewee’s discourse.

6. Results

The quantitative outputs reported in Figure 1, Figure 2 and Figure 3 are descriptive (frequencies and percentages) and are used to characterize the distribution of narratives and themes across outlets; they provide context for the subsequent qualitative interpretation.

6.1. Narratives of the Folha and Gazeta Portals About Vaza Jato That Focus on the Truth or Lack of It

Narrative 1—The messages published by TIB were true and there was no crime in their dissemination—Folha

From the very first publications, Folha reinforced the idea that the messages published by TIB were true (“it had access to the material and found no indication that it may have been tampered with”—23 June) and that the reporters dedicated a lot of time to “analyzing the dialogues, examining the context of the discussions in the various message groups and checking the information found to verify the consistency of the material” (23 June). Folha also clarified that it “does not commit an illicit act to obtain information, but may publish information that was the result of an illicit act if there is a public interest in the material found” (25 July) and that the confidentiality of its sources is constitutionally guaranteed.

The scandal Investigation into Glenn Greenwald and threats against the journalist was one of the most mentioned scandals (6.03%)—Figure 1.

This scandal referred to the possibility of an investigation into TIB’s editor-in-chief, Glenn Greenwald, who came to be suspected by Lava Jato defenders of being involved in collusion with hackers to illegally obtain conversations. Defense of publications based on the messages (6.69%) was among the main actions taken by the agents active in the Folha pieces—Figure 2. The framing Condemnation by the media of the actors involved in the scandals published in Vaza Jato as the second most common subcategory (16.60%) is indicative of this narrative.

Figure 1. Scandal in the spotlight in Folha’s news pieces (9 June to 9 September 2019). Source: Authors’ own elaboration.

Figure 1. Scandal in the spotlight in Folha’s news pieces (9 June to 9 September 2019). Source: Authors’ own elaboration.

Figure 2. Actions taken by the active agents in Folha’s news pieces (9 June to 9 September 2019). Source: Authors’ own elaboration.

Figure 2. Actions taken by the active agents in Folha’s news pieces (9 June to 9 September 2019). Source: Authors’ own elaboration.

Narrative 2—The TIB may have altered messages and may have committed crimes, which is why Glenn Greenwald needed to be investigated—Gazeta

Many of the positions of Lava Jato or Sergio Moro, which were highlighted, suggested that the messages could have been tampered with or forged, since it was not possible to check their authenticity. There were also allegations linking the crimes committed by the hackers to the Vaza Jato publications. In the publication “What is known so far…” (10 June), Gazeta referred to the messages as “alleged conversations”, claimed that TIB had published them without listening to those involved and stated that it was not yet possible to understand the involvement between the media outlet and the hackers. The narrative has intensified since the hackers’ arrests at the end of July. The scandal of the hacking of the authorities was one of the most recurrent—Figure 3.

Figure 3. Scandal in the spotlight in Gazeta’s news pieces (9 June to 9 September 2019). Source: Authors’ own elaboration.

Figure 3. Scandal in the spotlight in Gazeta’s news pieces (9 June to 9 September 2019). Source: Authors’ own elaboration.

Glenn Greenwald was one of the most cited active agents, in seventh place (4.02%). In addition, the action taken Allegations of illegalities in the production of stories based on leaked messages appeared in seventh place (5.53%).

The framing subcategory Accusations of illegalities in the production of stories (10.74%) and the subcategory related to allegations of unproven veracity and possible tampering with messages (13.42%) were the most common.

Narrative 3—Vaza Jato publications undermine the work of MPF prosecutors in the fight against corruption—Gazeta

There was a defense that the journalistic pieces were a campaign against Lava Jato to exonerate the corrupt politicians and to free Lula. The framing subcategory Damage to the fight against corruption (17.79%) is one of the most common. Along these lines, Deltan Dalllagnol said in an interview with Gazeta that there would be a movement orchestrated by the executive, legislative and judicial branches against corruption (23 August). Speeches in favor of Lava Jato were also echoed by Jair Bolsonaro’s ministers and other judicial and political actors. It was a narrative also present in Folha, but with less emphasis.

Narrative 4—Lava Jato agents’ partiality and inconsistencies—Folha

This narrative focuses on experts who agreed that Sergio Moro had been partial and had made defense allegations that contrasted with the law. Several judges and legal experts, including Supreme Federal Court justices, were heard by the media outlet and agreed with the possibility of suspicion. The narrative also exposed contradictions between the words and actions of the Lava Jato prosecutors. It was stressed that, from the beginning, the prosecutors did not deny the authenticity of the conversations and that “four days later the discourse of the Lava Jato task force changed and possible adulterations were pointed out” (13 June).

6.2. Discourse in the News and Editorials of the Folha and Gazeta Portals on the Value of Published Information About Vaza Jato

From the start, Folha’s journalistic pieces were dedicated to reproducing the entire dialogues published in TIB’s reports on Vaza Jato and assumed them to be true and relevant to be covered. Through its partnership with TIB, Folha published exclusive news pieces based on the database of leaked messages. As a result, a complete exposition of the messages referred to in the texts is built up so that the reader can have a broader context and anchors for interpretation (e.g., “Read excerpts from messages…”—30 June; “Understand the Lava Jato episodes…”—10 June; summaries).

The strategy of legitimacy and positioning in favor of the authenticity of the leaked conversations is used in the publication that begins Folha’s own reports (e.g., “The reporters searched for names of Folha journalists and found several messages that these professionals exchanged with members of the task force in recent years, thus obtaining a strong indication of the integrity of the material”; “The report detected no indication that it may have been tampered with”—23 June).

The editorials also defend the authenticity of the conversations and the right of TIB and other media outlets to publish the leaked conversations (“evident public interest justifies publication”—27 July). It is also pointed out that Folha “found no signs of tampering when examining the messages” (27 July). An attempt is also made to rebut Moro’s and Dallagnol’s arguments against the messages (“the argument that tries to question the reliability of the messages is fragile”—27 July). The editorials also adopt a critical tone towards the defense’s strategy of questioning the veracity of the conversations (“There was no denial among the participants in the conversations”—21 June; “the defense’s strategy is diversionary and probably innocuous”—10 August).

At Gazeta, the news discourse focuses on the official notes from the Ministry of Justice and the MPF in the title and lead of the publications (“Lava Jato denounces hacker attack against prosecutors”—9 June; “Lava Jato task force questions veracity of messages”—30 June), leaving the content of the messages in the original materials in the background. Unlike Folha, there is no publication of excerpts from the dialogues in headlines, bylines or leads.

In the first news pieces, the repercussions of the messages released by TIB were concentrated in a single news piece (“What is known so far…”—10 June) which was constantly updated until 9 July. Since then, with the publication of reports by Folha, Veja, BandNews FM and El País, some of the scandals have received exclusive space in the media (“Lava Jato considered seeking evidence against Gilmar Mendes…”—6 August).

The journalistic pieces place Moro and the Lava Jato task force as the main agents of the actions and that question the veracity and authenticity of the messages from the beginning of their disclosure (“Moro denies authenticity of conversations…”—15 June; “Lava Jato calls “fake news”…”—22 June). By echoing various excerpts from the official memos, the official discourse of Moro and the prosecutors is reinforced, leaving room for strong criticism from the prosecutors (“Use and editing out of context to support accusations and distortions that do not correspond to reality”—6 August). As a strategy for positioning distrust in relation to the authenticity of the messages, it is reinforced that those involved “deny their authenticity” (10 June) and suggests “signs of editing” (29 June), “fake news” (22 June) and that the dialogues may have been “forged” (12 June). The seriousness of leaks “of documents and data on ongoing strategies and investigations and on the personal and security routines of the members of the task force and their families” (9 June) is also highlighted.

Gazeta’s news pieces frequently use the term “supposed” for messages and dialogues and the future tense (“would have written”; “would have stated”; “would have informed”—10 June). These linguistic choices are justified because TIB “did not expose the primary information as copies of the screenshots and audios” (“All the information published so far is the result of editing or interpretation of the data set that the site claims to have obtained”—10 June). Without publishing the conversations in full, Gazeta argues that there could be a risk of “misrepresenting facts”, “falsifying information” and “spreading fake news” (10 June).

It is highlighted the hacker attacks that prosecutors and politicians have suffered, the “criminal action”, the “serious and illicit affront to the State” (10 June), as well as the “sensationalism and criminal violation of privacy” (18 July), trying to associate the hackers with TIB journalists, undermining their credibility. TIB’s “left-wing editorial line” and Glenn Greenwald’s connection to a left-wing politician (“married for 14 years to David Miranda”—10 June) are emphasized.

The editorials criticize the publication and repercussion of the messages (“There is still no proof that the messages exchanged were actually sent by Moro and Dallagnol”—11 June; “there is no evidence so far that the conversations are true”—29 June). The crime of hacking the authorities is emphasized (“there is no doubt that there was a crime on the part of those who hacked the authorities’ cell phones, the possible mastermind of the hacking and those who distributed them to The Intercept website”—11 June), even though the TIB is exempt from committing crimes (“they are exercising their right to freedom of the press”—11 June).

There are various criticisms of the accusations against Moro (“absurd”, “unfounded” and “not consistent”—11 June). There are accusations that the left, which is against Lava Jato, was the main beneficiary of Vaza Jato. This raises suspicions about the journalists’ real interest in publishing the reports in question (“a new attempt to impose the fable that Lava Jato is nothing more than a conspiracy to bring down the PT [Workers’s Party] and put Lula in jail”—11 June).

6.3. Positioning of the Editors of the Folha and Gazeta Portals That Motivated Their Decisions on Whether or Not to Publish Information About Vaza Jato

Folha was “the first media outlet sought out. They [TIB] were after our credibility” (Sergio Davila). Davila points out that the partnership involved the credibility of Folha and that of Greenwald, “a renowned journalist, who had given a world scoop”, known worldwide for having produced reports with data leaked by former Central Intelligence Agency (CIA) agent Edward Snowden.

Because of the length and the publication potential of the material, journalistic production on the subject became, “in political coverage, a priority content, it was our main agenda. So much so that we deployed this team of four reporters, which is unusual”. The interviewee emphasizes Folha’s control over the material obtained (“Folha would control everything it would be published, be edited and investigated by our team. The documents would not be published uncritically”), including the deadlines for publication. “Our team would investigate and when they thought there was something consistent, they would publish it. Without worrying too much about timing, we only published when it was safe to do so”. All the material was obtained through primary sources, i.e., access to the original materials obtained by the hackers (“It was nothing second-hand, it was all first-hand”).

Davila revealed a concern with producing more contextualized reports so as not to use only the leaked messages as the only elements (“the material would serve as a starting point for an investigation and from that investigation we would publish the reports”). Davila underline that, “during Vaza Jato, we contextualized, listened to experts, answered questions about whether any laws had been broken or not”.

The editor-in-chief reinforces that Folha’s ideals of independence were decisive for the proposal of the partnership (“Glenn identified with the newspaper’s position of independence, with the editorial project of critical and non-partisan journalism”) and for the newspaper to maintain its credibility in a diachronic line, even after the criticism of the partnership having been made with a left-wing website (“Part of the public questioned whether this would be a shift of Folha to the left. I did not worry about this kind of questioning at the time because I know the newspaper we make and that it would quickly prove that independence would continue”). There was a concern to maintain independence, but this was not “a deliberation, an order, but something that was naturally proven on a daily basis”. According to the interviewee, other, external journalists questioned the partnership. “My response was: wait and see if independence is compromised or not. The years have shown that it is not”.

Schenkel stresses that there are doubts to this day because “nobody has confirmed the authenticity of the messages”. “I was not going to publish it because you could also be manipulating something. If I did not have the raw material or a recheck, it would not work”). This is the “primary source” argument. “If the hacker has the ability to get hold of this documentation, he may have put a point in the wrong place that changes the interpretation. How can one be certain that the hacker did not alter the content?”.

The interviewee also criticizes the production of the reports and the use of the conversations leaked by TIB. “There were several assertions based solely on their credibility. The Intercept was a very new media outlet with a history of very big mistakes. It is different from Folha in that we know its history”. Regarding TIB’s journalistic performance, the interviewee criticizes the quality of the text which “is very confusing, very editorialized, they mix opinion with information, make a cut, change things in order”), the stance adopted by the reporters (“they called for the arrest of Lava Jato agents”) and the mistakes made by the website in covering other cases (“culpable rape”; “Winner case”), including in relation to a report on Gazeta itself (“article in 2018 full of errors”).

With regard to Gazeta’s coverage of Vaza Jato, Schenkel explains that, because he did not have access to the full dialogues, he was limited (“We were tied down. How are you going to do in-depth material on something you do not have?”). That is why the subject did not become a priority in the newsroom. One journalistic solution was to highlight other points of view on the repercussions of the scandals, such as those of the Lava Jato prosecutors, Sergio Moro and other government authorities, who “are public officials and you can verify what they said. So it is natural that we feel more comfortable publishing the repercussions”. In this way, “Gazeta showed another side of Vaza Jato, which was not more prominent in other media. That is why the public would come to Gazeta to get access to this information”.

7. Discussion

The comparative analysis of Folha and Gazeta shows that similar sets of leaked materials were narrated through different discursive strategies—emphasizing verification and partiality in the first case, and suspicion and damage to anti-corruption efforts in the second. These contrasts highlight how newsroom decisions cannot be fully understood in isolation from the broader institutional and structural context in which Brazilian journalism operates. To explain these differences more systematically, we now outline a set of operational mechanisms—grounded in the literature on Brazilian media and democracy—that link media system structures with newsroom practices.

Our findings can be explained by three interrelated mechanisms grounded in scholarship on Brazilian journalism:

(1)

Editorial instrumentalization under concentration/polarization. Ownership concentration and polarized political competition condition how legacy outlets frame judicial controversies and leaked material, often heightening caution about becoming actors in partisan disputes (Porto, 2012). This dynamic reflects a long-standing pattern whereby media conglomerates maintain close ties with political and economic elites, which historically limited the press’s autonomy as a watchdog (da Silva, 2000).

(2)

Professional verification norms and trade-offs with illicitly obtained leaks. Legacy newsroom routines privilege corroboration and provenance, constraining what, when, and how to publish even when documents appear newsworthy (Waisbord, 2013). While professionalism and objectivity expanded significantly after redemocratization—especially in outlets like Folha—these gains have often coexisted with commercialization pressures, denuncismo practices, and sensationalism that complicate the boundary between investigative rigor and partisan exposure (Matos, 2006).

(3)

Digital-native investigative entrepreneurship. Although not a case in our sample, TIB’s trajectory illustrates a field position characterized by higher risk tolerance, digital visibility, and exposure to harassment—factors that help explain why different outlets arrive at divergent publication strategies (Global Investigative Journalism Network, 2019). This resonates with broader debates on media democratization in Brazil, where the weakness of public service broadcasting and the persistence of authoritarian legacies leave limited institutional support for independent journalism, reinforcing the role of entrepreneurial and digital-native actors in expanding the public sphere (Matos, 2013).

Articulating these mechanisms clarifies how journalism operates in Brazil’s media system and democratic context, connecting structural concentration and institutional legacies with newsroom practices and editorial strategies. It also helps to account for the contrasting narratives we document between Folha and Gazeta, highlighting how professional norms, structural constraints, and field positions intersect to shape divergent responses to the Vaza Jato leaks.

Declaring the digital era a “golden age for journalism”, the director of the International Consortium of Investigative Journalists (ICIJ) Gerard Ryle, says that technology is allowing information to leak on a scale that could not have been imagined, enabling journalists to obtain incredibly detailed information and stories that they would not have been able to before (Posetti, 2017, p. 104). This is the case with Vaza Jato, which, thanks to leaks, was discovered and, thanks to journalists, made known to the world.

One of the narratives constructed by Folha focuses on the veracity of the messages and, therefore, their potential to be disseminated. The veracity of the messages was made possible thanks to first-hand access to the material, i.e., as a primary source, analysis of the leaked dialogues, examination of the context and confirmation of the information. All these conditions and tasks are required of quality journalism. As Ireton and Posetti (2018) reminds us, “the discipline of verification” “separates professional journalism from the rest” (p. 19).

In the Brazilian case, journalists would not be committing a crime by publishing stories that come from illegal means. According to Brazilian law, the crime happens the moment someone violates the privacy and has illegal access to conversations, and not at the moment the journalist disseminates what he has known (as long as the journalist is not involved in the hacking). However, in the British context, The Guardian said that evidence from British intelligence documents revealed by Snowden also showed that a GCHQ (the UK’s intelligence, security and cyber agency) information security assessment had listed investigative journalists as a threat along with terrorists and hackers (Ball, 2015).

The Folha journalists evoked the public interest as a motive for publication. But journalists need protection, in terms of defending the public interest, recognized by the law and the courts (Posetti, 2017, p. 26). If journalists are not guaranteed this legal and judicial safeguard, there could be a retreat from the possibility of publishing content in the public interest. Many countries, for example, in the Pacific and Asia, criminalize journalists who publish leaked information (Posetti, 2017, p. 67). One of the great fears about Vaza Jato was that Glenn Greenwald, editor-in-chief of TIB, could have been prosecuted or imprisoned for publishing leaked conversations.

One of Gazeta’s narratives is the suspicion that Greenwald may have tampered with messages, which would constitute a crime. This narrative focuses on Greenwald as an anti-hero or villain, contrasting with Folha’s narrative which presents the disseminators of the content (journalists) as legitimate executors of their profession.

Another narrative (Gazeta) exposed the damage caused by the publication of Vaza Jato in the fight against corruption and Lava Jato, arguing that the rationale behind the publication was to exonerate corrupt figures. On the other hand, another narrative (Folha) shows the partiality and inconsistencies of the Lava Jato agents. We find a Manichaeism between the two narratives in the opposition between us (good) and them (bad).

The analysis of Folha’s news and editorial discourse and the interview with its editor-in-chief also showed the ethical care taken to confirm that the information published was true, complying with the precepts required of a full democracy (e.g., Knight Foundation, 2023; Michailidou et al., 2023). Gazeta’s news and editorial discourse gives greater prominence to Moro and the Lava Jato task force, which questions the veracity of the messages. In this way, Gazeta reinforces the official discourse, as it can check with public sources what they have said, and questions the truth of the leaked messages. In a Manichean line, the speeches of Moro and the Lava Jato agents catalog hacker attacks as criminal, serious and illicit. In the editorials, there is an interpretation that the main beneficiary of Vaza Jato is the Brazilian political left, which seeks to highlight a certain promiscuity and collusion between journalistic and political power. In his interview, the editor-in-chief of Gazeta assumes that Gazeta is a favorable media outlet and defender of Lava Jato. This position may raise doubts about its influence on Gazeta’s journalistic approach adopted in relation to Vaza Jato. Hiltunen (2022) finds that political (e.g., left versus right) and social polarization are symptoms and driving forces behind the erosion of public trust in journalism. In this context, Horner et al. (2021) show that individuals are more likely to believe headlines that align with their existing beliefs. This means that individuals politically identified with the right would be more likely to believe the view that the publication of Vaza Jato favors the left.

In face of the lack of certainty regarding the veracity of the information, Gazeta, given that it did not have access to the leaked material as a primary source, chose not to be a promoter of possible disinformation and not to contribute to the continued influence effect (Ecker et al., 2022), associated with a possible correction, and even to a breach of trust on the part of the public. Folha, being certain about the veracity and having access to the leaked conversations as a primary source, chose to do what most U.S. journalists would do, that is, report on the statement because it is important for the public to know about rather than not report on the statement because it gives attention to the falsehoods and the public figure (Gottfried et al., 2022).

There are no factors present that motivate the consumption of disinformation, such as the scarcity of own reports (e.g., Brennen, 2020; Hendrickx & Van Remoortere, 2023), given that Folha used a team of four journalists—“which is not common” (Davila) and shows the adoption of the management approach (Banerjee et al., 2023) and diverse presenting reporting team (Watson, 2024a) to build trust with the public –, producing several independent reports. Also to build this trust, better explanations of complex stories (e.g., “Understand…”, “What is known so far”) and simpler language/accessible formats (e.g., summaries) are practiced.

8. Conclusions

This conclusion synthesizes the study trajectory. We began with a focused comparative question: how do two newsrooms with different field positions and access conditions—primary access versus no access—make publication decisions about contested leaks? To answer it, we assembled a corpus of 828 items from the first three months after the initial disclosures and applied three complementary techniques—narrative analysis of the full corpus, critical discourse analysis of a purposive subcorpus of news and editorials, and elite interviews with the two editors-in-chief. These steps yielded comparative findings: Folha de S.Paulo emphasized verification and partiality; Gazeta do Povo emphasized suspicion and damage to anti-corruption efforts. We then interpreted these results using three operational mechanisms that link media-system structure to newsroom practice: (1) editorial instrumentalization under concentration/polarization; (2) verification norms and trade-offs with illicitly obtained material; and (3) digital-native investigative entrepreneurship and its field effects.

Taken together, the results highlight that the decision to publish or withhold leaked content is shaped not only by access conditions but also by newsroom identity and field position. At the same time, they suggest that transparency about verification routines, contextualization, and editorial independence are central to reducing disinformation risks in similar cases. Based on these findings, several recommendations can be made to avoid disinformation associated with the publication of leaked content, such as disseminating the content if there is a guarantee of authenticity and obvious public interest; making a complete exposition of the messages; building a broader context and anchors for interpretation; using clear language and accessible formats; publishing own reports; highlighting the denial or non-denial of those involved in the leaked messages; if the veracity of the information is confirmed, the journalist can transpose excerpts from the dialogues into headlines, bylines or leads; when in doubt, using the word “supposed” and the future tense of the compound past tense; monitoring what competing media outlets are doing and saying (news value of competition); in the case of partnerships, establishing them with credible media outlets; not publishing content uncritically; controlling everything that is published; listening to experts; deciding while respecting the newspaper’s editorial identity; demonstrating the newspaper’s independence; considering the argument of the primary source; marking the boundary between information and opinion; showing another side of the case/scandal to attract the public to the exclusive information; knowing and deciding considering the ethical and legal framework in which the journalist operates. Despite these contributions, the study has important limitations. First, it focuses on two outlets chosen for their contrasting positions, which means the findings are analytically rather than statistically generalizable. Second, the temporal scope is limited to the first three months of the Vaza Jato disclosures, leaving longer-term developments outside the analysis. Third, the evidence from elite interviews reflects the perspectives of two editors-in-chief and may not capture dissenting voices within each newsroom. Fourth, the statistical analysis is descriptive rather than inferential, aimed at characterizing patterns but not establishing causal relations. Finally, the study is bounded linguistically and geographically, concentrating on Brazilian Portuguese outputs of two outlets and not considering cross-platform dynamics such as closed messaging applications.

In light of these limitations, the study nonetheless makes contributions at three levels. Academically, it advances operational theory by specifying mechanisms that connect system structure to publication choices under uncertainty. Professionally, it sheds light on narrative and editorial discourses and on the rationales of gatekeepers, while also offering practical cues for handling leaks. Socially, it contributes to transparency (Banerjee et al., 2023), which is a necessary condition for public trust in the media. Future research should broaden the sample of outlets and extend the time horizon to examine whether the mechanisms identified here are consistent across other leak-driven controversies and whether newsroom practices evolve over time.