News or not news? Information sources in a nationalist WhatsApp group during Chile’s 2019 protestsAbstract
This article analyzes links shared in a nationalist public WhatsApp group during an intense social outbreak in Chile in 2019. Through content analysis we assessed the quality and credibility of sources shared in the group. The findings reveal: 1) a prevalence of a subjective vision of the events; 2) a prevalence of thematic YouTube channels and personal profiles; 3) topics with critical views on migration, politicians, and the protests; and 4) an abundance of second-hand decontextualized information, and broken links. We discuss the limitations and implications of the results for democratic societies.
Keywords
Chile; WhatsApp; YouTube; news quality; nationalism; homophily
Resumo
Esta pesquisa analisa os links compartilhados em um grupo público de patriotas no WhatsApp durante a revolta social chilena de 2019. Aplicando análise de conteúdo avaliamos a qualidade e a credibilidade das fontes compartilhadas no grupo. Os achados revelam: 1) prevalência da visão subjetiva dos acontecimentos; 2) prevalência de canais de nicho e perfis pessoais no YouTube com temas específicos; 3) visões críticas sobre migração, políticos e protestos; e 4) abundância de informação de segunda mão descontextualizada e links quebrados. Discutimos as limitações e implicações dos resultados para as sociedades democráticas.
Palavras-chave
YouTube; WhatsApp; Chile; qualidade das notícias; nacionalismo; homofilia
Resumen
Esta investigación analiza los enlaces compartidos en un grupo público de patriotas en WhatsApp durante el estallido social chileno de 2019. A través del análisis de contenido evaluamos la calidad y credibilidad de las fuentes compartidas en el grupo. Los hallazgos revelan: 1) prevalencia de la visión subjetiva de los eventos; 2) prevalencia de canales específicos de temas de YouTube y perfiles personales; 3) visiones críticas sobre la migración, los políticos y las protestas; y 4) abundancia de información de segunda mano, descontextualizada y enlaces rotos. Discutimos las limitaciones y las implicaciones de los resultados para las sociedades democráticas.
Palabras clave
YouTube; WhatsApp; Chile; calidad de las noticias; nacionalismo; homofilia
Résumé
Cette recherche analyse les liens partagés dans un groupe WhatsApp public nationaliste lors de l’éclosion sociale au Chili en 2019. Grâce à l'analyse du contenu, nous avons évalué la qualité et la crédibilité des sources partagées dans le groupe. Les résultats révèlent : 1) la prédominance d'une vision subjective des événements ; 2) la prévalence des chaînes YouTube thématiques et des profils personnels ; 3) des sujets avec des opinions critiques sur la migration, les politiciens et les manifestations ; et 4) une abondance d’informations de seconde main décontextualisées et de liens morts. Nous discutons des limites et des implications de ces résultats pour les sociétés démocratiques.
Mots-clés
YouTube; WhatsApp; Chili; qualité des informations; nationalisme; homophilie
Introduction
During political and social crises, people have a greater need for orientation, (Weaver, 1980) which increases pressure on both official and alternative media outlets, and can lead people to seek alternative information sources. This impulse, however, is not necessarily accompanied by a greater capacity to discern between legitimate and illegitimate information sources, between opinion and fact, or between real content and fake news, to mention just a few related problems.
As we will see in the case studied, a major social crisis in Chile, the convergence of factors such as the urgent need for information, the crisis of faith in traditional media and the absence of digital skills in contexts marked by a great diversity of sources (Van Aelst et al., 2017) create a perfect storm for the dissemination of information of questionable quality. The growing use of interpersonal communication platforms as information sources generates environments that are more protected from public scrutiny. There is an important distinction between these and other platforms, such as X, Instagram or public Facebook groups, on which the great majority of the information is accessible. On those more public platforms, independent fact-checking agencies can verify content, either independently or financed by social media companies such as Meta7, in response to social criticism and even threats of legal action over the circulation of false content, as occurred in Brazil with Telegram during the 2022 presidential elections (Galarraga, 2022).
This article problematizes the growing role of WhatsApp as an information source in the Global South. We do so by analyzing the links shared in a public Chilean WhatsApp group with an extremist, nationalist political orientation during the most important social crisis in the country in the past few decades, which was known as the estallido social — the social outburst. We approach the object from a case study perspective because it allows conducting an in-depth empirical inquiry about a contemporary phenomenon within its real context (Yin, 2018). We also concur with three of the arguments that Yin (2018) makes for selecting this approach: 1) our main research question is related to how a group of nationalists inform themselves on WhatsApp during a social crisis; 2) we have no control over their behavioral events; 3) the focus of the article is a contemporary phenomenon. Despite criticisms of single case studies, and being unfit for a traditional generalization of the findings, we agree with Fidel (1984) that the case study research allows generating relevant findings beyond the individual cases. Based on this analysis, we offer a necessary initial approach to news habits in contexts of political emergency in spaces that have received limited attention in the literature, as is the case of mobile instant messaging services (MIM).
Need for orientation and Chilean media during a major social crisis
In Chile, the second half of 2019 was marked by what came to be known as the estallido social, a series of massive protests triggered by a 30-peso increase in the subway fare (equivalent to US$ 0.04 at the time, about 3.75%). The protests began in Santiago the week of Monday, October 14, but quickly had national repercussions and continued through at least mid-November. The most important date related to the uprising was October 18, 2019 (later known as 18O), when supermarkets, buses and even churches were looted and police and subway stations set on fire (Santiago […], 2019). Over 30 people died, and some 400 people suffered eye injuries (Amnesty International, 2020), during one of the longest, most violent, and important protests in Chilean history, possibly the most important (Salazar, 2019).
An important and generalized social crisis increases levels of uncertainty, triggering high levels of what is called a “need for orientation” (NfO) (Weaver, 1980), a psychological arousal construct related to an urge to fill information gaps and help navigate our surroundings (McCombs; Weaver, 1973). While two dimensions are often used for studies on the NfO, framework-relevance and uncertainty, which were present in the context of the estallido, a third variable was operationalized in the foundational social psychology research of Jones and Gerard: the likelihood that a reliable source of information is available to shed light on topics of interest (McCombs; Weaver, 1973). This variable was at some point “taken as given” since it was considered that in the mass communication society, all citizens had easy access to reliable information sources. But in recent decades citizens quickly migrated from environments of information scarcity, moderated by gatekeepers, to environments characterized by an information overload (Santos; Valenzuela, 2022) moderated by algorithms and individual source selection criteria, amidst an important crisis of trust in legacy media. The diversity of sources has not translated solely into multiplicity of viewpoints but also in an informational environment in which it is difficult to discern reliable sources from partisan ones, propaganda, misinformation or simply low-quality content (Lewandowsky; Ecker; Cook, 2017).
Chileans were placed under a curfew beginning the night of 18O — which lasted a little over a week, at least in Santiago (Chile […], 2019) — meaning that social media and instant messaging apps were perhaps the only options for discussing and sharing opinions about the events. At that time, June 2019, Chile had 72% Internet penetration (Newman et al., 2020), 134.5 cell phones per 100 inhabitants — that is, significantly more devices than people — and 22 million Internet accounts including mobile internet and cellular service (Subtel, 2019). Eighty percent of the population used WhatsApp and nearly 1 in 3 said they got news from this platform (Newman et al., 2022).
Moreover, press coverage of the protests was not well-received by the citizens. This resulted in an important drop in trust (by approximately 15%) in traditional media outlets between 2019 and 2020 (Newman, et al., 2020). Luna, Toro and Valenzuela (2022) found that while traditional and alternative media experienced an increase in the number of “shares” on Facebook, the increase for traditional media was of a “pulse” type of shock, and soon returned to previous standards. The increase for alternative media was a “step” type of shock, and the increase was sustained for nearly one year after the protests.
MIM platforms present significant barriers for the academic community or journalists, such as access to content for verification, making the work of analyzing the information that circulates on such platforms and their social effects more difficult. However, they cannot be ignored in a country like Chile, for the reasons described above. For this reason, we believe that a valuable first step is to describe various dimensions of communication patterns of users of the group studied, such as the sources of information shared in a situation of increased NfO. Thus our first research question is:
RQ 1: Which information sources were shared by members of the nationalist WhatsApp group during the Chilean social crisis?
Homophilic communities and motivated reasoning
The idea of creating communities has been present throughout human evolution, but some types have acquired more strength lately due to the individualism and alienation that characterize modern society. Bauman (2003) has stated that a community can serve as a refuge from a hostile world. Groups develop around identity axes that range from topics like food (Kozinets; Patterson; Ashman, 2017) to ideology. The dynamics and purposes of communities are not static but react to social events and adjust based on milestones. Tönnies (2009) distinguishes between ‘community’ and ‘society,’ noting that the former has no awareness of its perpetuation in civilization and only exists with its current members. By contrast, the concept of ‘society’ is more complex. This differentiation is useful for understanding levels of social trust among different groups (Fukuyama; De Salcedo, 1996), but it is also interesting for categorizations like those of WhatsApp groups which become digital communities.
Given the characteristics of a group related to an identity marker as potent as nationalism, one might expect that a social crisis like the estallido would be a topic of interest, increasing the need for orientation. However, groups with a high level of ideological identification are more predisposed to seek out strong alignment based on values and beliefs — a phenomenon called ‘partisan selective exposure’ (Stroud, 2010) — without concern for journalistic quality or source veracity to reduce cognitive dissonance (Stroud, 2010). In other words, due to a desire to reduce the psychological discomfort of facing statements that go against their existing beliefs or convictions (Festinger, 1993), people choose consonant sources (that are in keeping with their beliefs), avoiding exposure to the conflict caused by dissonance (something that goes against their beliefs).
The idea of selective exposure is closely linked to the concept of motivated reasoning (Fischle, 2000), which is defined as ‘the psychological phenomenon by which individuals' biases and predispositions influence their subsequent attitude formation’ (Kopko et al., 2011, p. 273). Lodge and Taber (2000) state that there are four ‘ideal’ types of reasoning depending on whether strong or weak objectives of direction and precision are sought. As such, these authors call a thinker who presents weak precision objectives (that is, who does not spend a lot of resources on ensuring the quality of the information found) and high direction objectives (that is, a clear confirmation bias) is called a ‘partisan reasoner’. This concept can be applied to the nationalist WhatsApp users who share information in the group that is the focus of this article. Nir (2011) found that this type of reasoner is likely to overestimate the value of information sources that confirm what they already knew, while underestimating the value of opinions that go against thoughts that have already been generated. As such, a partisan reasoner is someone who focuses more on the content than the quality of the information or its source.
Source quality and credibility
The literature on the quality of information sources points to a consensus associated with a half-dozen variables related to the provision of information that does not distort reality and is not second-hand (Alessandri et al., 2001; De Pablos; Mateos, 2004; Gutiérrez-Coba; Salgado-Cardona; Gómez-Díaz, 2012). These are: usefulness, understood as a quality of information that helps people with their daily needs; source diversity; originality, defined as the existence of topics unique to the medium; impartiality, measured by the presence of diverse and balanced perspectives; contextualization, which is additional information that provides background information about the event described; and clarity, which refers to content that is legible in terms of spelling, punctuation and grammar.
These are elements which, depending on how they are deployed, should give the published information more or less credibility. In other words, they relate to trust in the rigor, truth and precision of what is published (Flanagin; Metzger, 2000), based on the expertise and reliability that can be attributed to the source. As such, information credibility is a concept in tension in today’s media reality. On the one hand, the main factors that produce it tend to be the media outlet’s previously constructed reputation and/or that of the author of the information, and simultaneously the identification with the editorial stand (Gutiérrez-Coba; Salgado-Cardona; Gómez-Díaz, 2012). Thus, a long-nurtured reputation of legacy media (Hall, 1992) becomes an important intangible asset, as news organizations’ reputation impacts news sharing behavior and other factors (Karnowski et al., 2020). However, this asset is in crisis in the age of social media, given that peer evaluation and perception of quality are relevant factors for users in their determination of media reputation and credibility (Aruguete; Calvo; Ventura, 2023) and that long-standing reputation may drown in a sea of peer-created likes and shares. More than ever, those who consume information thus form an idea of credibility that goes beyond the intrinsic quality of the content.
Additionally, the maintenance and updating of content make for an indicator that is more frequently utilized for that assessment because less reliable sites tend to contain inoperative (unavailable) links or links that are replaced by more current pages. A site or channel that is not stable over time — on its own or in terms of its links — should set off alarms for users (Oliván; Ullate, 1999).
Based on these issues, we ask:
RQ 2: How is (a) the quality and (b) the credibility of information sources shared on the nationalist group?
Social media and WhatsApp as information sources
Initially, the advent of digital media was seen to increase source plurality and diversity. It could be seen as a positive space for a more participatory (Shirky, 2009) form of democracy (Wolton, 2010; Zuckerman, 2014) in which products of collective intelligence are harvested (Lévy, 2004). However, this vertiginous move from an environment of information scarcity to one of information overload (Santos; Valenzuela, 2022) brought about a complex situation in which finding reliable, high-quality information about politics is not easy (Lewandowsky; Ecker; Cook, 2017). The diversity of options can create a news overload that instead of assisting can impede decision-making (Bawden; Robinson, 2020). In addition, according to Van Aelst et al. (2017), diversity can lead to greater fragmentation and polarization of political positions, and encourage ‘post-truth-like’ relativization of facts and increase inequities in political knowledge.
Instant messaging applications were originally created for interpersonal communication, but they have become communication tools with many other functionalities, including group communication, ‘bots’ with automated responses and dissemination channels that are more massive and unidirectional in nature. The way that this kind of tool is used has become more complex and incorporated other activities beyond daily conversation. The same is true for WhatsApp, the world’s most popular MIM application, with over 2 billion users at the time of the writing (Statista, 2022).
It is therefore not surprising that these new opinion shaping spaces host discussions of political issues of national interest, given their hybrid nature, which blends private and public affairs (Valenzuela; Bachmann; Bargsted, 2021; Treré, 2020). As Milan and Barbosa (2020) affirm, “WhatsApp merges into a single digital platform the private and intimate sphere (of interpersonal exchanges) and the public realm (of group interactions) ‘integrating’ the vernacular and the political”. For the group studied, WhatsApp is a tool for political discussion and activism (Gil de Zúñiga; Ardèvol-Abreu; Casero-Ripollés, 2021) and a safe space for sharing political positions that could be controversial in other fora (Valeriani; Vaccari, 2017) or create problems for users in contexts of censorship (Johns, 2020). However, recent studies suggest that the use of social media or MIM for news consumption can also create problems. These include a false sense of being informed (Gil de Zúñiga; Weeks; Ardèvol-Abreu, 2017), exposure to more politically radical news (Valeriani; Vaccari, 2017) and even less knowledge about public affairs (Valenzuela; Bachmann; Bargsted, 2021), which are problems that can be amplified due to a strong political group identity (Saldaña; McGregor; Johnson, 2021). Recent studies in Brazil and Chile show that highly active users may fall into the paradox of participation: the more engaged in political talk and the more participative they are, the more likely social media users are to share misinformation (Valenzuela et al., 2019; Rossini et al., 2021).
However, all the studies mentioned above are based on surveys, that is, they are self-reported, and do not provide a window to the actual content that is being shared. One of the main contributions of this exploratory study is, thus, through an initial approach with a content analysis of empirical data, to understand the patterns of information consumption and link-sharing by the members of a single WhatsApp group and other nuances of the informational behavior of the community of nationalists during the social outbreak in Chile. Sharing links is a common activity among users of instant messaging groups, which became “meso news spaces” where users feel free to create or publicize news in a more controlled environment (Tenenboim; Kligler-Vilenchik, 2020). Actually, the main motivations for sharing news in spaces like WhatsApp are to avoid more public exposure, which is often perceived as uncomfortable, and to share specific interests with one or more recipients (Kalogeropoulos, 2021). Political news is a predominant type of content in this activity, especially in a politicized group, and its sharing among trusted individuals allows users to become informed, criticize, or express hatred regarding a public figure or current event, often as a form of catharsis (Lottridge; Bentley, 2018).
The importance of studying different aspects of public WhatsApp groups has increased in recent years because the platform has invaded various areas of political life, enabling information on politics to reach groups and people who would probably not receive it otherwise (Santos et al., 2019). WhatsApp also has been used to coordinate protests and other forms of political participation (Treré, 2020), discuss political actions (Milan; Barbosa, 2020), and avoid censorship during elections in countries with authoritarian governments (Johns, 2020). Telegram, WhatsApp’s main competition, has become the main means of communication for dissidents in countries in which there is significant control over communication apparatuses, such as Iran (Kargar; McManamen, 2018). Even when Russian authorities blocked Telegram, it continued to be used by the general public (Santos; Saldaña; Tsyganova, 2021), members of Putin’s opposition and, oddly, even government spokespeople to report to their followers (Salikov, 2020).
However, these political uses of applications designed for other purposes are not free of controversy. Encryption, security, and lack of public scrutiny have led users to adopt MIM platforms to spread hate and coordinate extremist activities by groups such as the Islamic State (Karasz, 2018) and the 2021 Capitol Riots in the US (Stewart et al., 2023) or to systematically spread misinformation (Ozawa et al., 2023). The main problem that this article considers is that the context of trust that develops in a very politically cohesive group like that of these nationalist WhatsApp users can have important implications at the level of political knowledge, particularly considering the possible absence of discernment between more or less established sources with varying levels of reliability. In a situation of crisis in which the need for orientation increases, the possible intensification of intra-group confirmation biases can lead users to ‘become entrenched’ in their information sources, increasing or concentrating their levels of exposure to politically more radical positions that may be self-censored on public-access social media such as Twitter (Valeriani; Vaccari, 2017). Research question 3 emerges from this discussion, in the understanding that descriptive data is useful for advancing in an understudied field such as that of the content shared within MIM extremist groups:
RQ 3: What are the dynamics and communications patterns that develop around the links shared in the nationalist group?
Method
In this article, we analyze data extracted from a WhatsApp group focused on Chilean nationalism. The name of the group is related to the idea of ‘patriotism’,8 and although the group existed for months prior to the social outbreak, we study its members’ reactions and behaviors regarding informational habits in the days leading up to the mass protests, during them, and in the days following 18O.
The link to access the group was found on a Chilean website with a WhatsApp group directory. Admission of new members to the group was not moderated by group administrators. This allows inferring that there was no focus on privacy but instead an effort to seek out more participants to disseminate the group’s ideas and discuss topics of interest. However, following the principle of maximizing protection of sensitive information (Ess; Hård af Segerstad, 2019 apud Franzke et al., 2020), in this case political information, the appropriate ethical steps were taken, as we will describe in the next section Data Collection and Treatment.
The research team chose a mixed-methods approach that iteratively combined qualitative and quantitative perspectives, with each informing the other. The initial analysis began from a qualitative perspective that allowed us to identify topics of conversation, reactions and external sources that were shared to identify categories of analysis and their metadata that could sustain the article. Additionally, all the links were seen, read, listened or watched in full (be it news, texts, images, videos etc.). This allowed the team to have a deeper understanding of the quality of the sources, as aspects to be used in the analysis. Using a quantitative approach, we proceeded to pre-process the data and analyze the content using the categories that emerged and some key metadata identified.
Data collection and treatment
The nationalist group was created in July 2019, and one of the researchers joined it in August of that year. The researcher did not identify himself or participate in the conversation because he was aware that the group members would most likely not welcome this type of observation. Given that it is a group with an extremist political orientation and the researcher’s orientation was different from that of other group members, the researchers could have faced threats or retaliation (Franzke et al., 2020) and/or lost methodological integrity (Bishop; Gray, 2017). Considering the epistemological approach and the affordances of the platform, we followed the best practices available in the discipline (Herrada Hidalgo; Santos; Barbosa, 2024) and decided to study the links shared within the group’s conversations.
A few months later, 18O broke out, bringing an unprecedented opportunity to understand how users react to information sources to make sense and act upon an exceptional social phenomenon. The researcher remained in the group until January 21, 2020, when he left it and downloaded the data (text only) through the application’s native functionality. The data were anonymized by associating an arbitrary numeral with each user based on their cell phone number, which serves as the identifier of the message author. No additional information was associated with the user. The original database is stored on an encrypted server and was not shared with the other researchers.
As the specific interest of this article is sources of information in the context of 18O, we limited the timeframe to five days before and five days after the event to observe changes in individual and group communication patterns related to 18O with regard to sources shared within the group. This yielded a corpus of 6,593 messages posted during the 11-day period selected: from 0:00:00 on 10/13/2019 to 23:59:59 on 10/23/2019. Heretofore we will call this the ‘period of interest’.
Content analysis
To analyze the content, the URLs were automatically extracted, which yielded 483 URLs posted during the period of interest. These were subjected to an initial review to build a codebook to support the content analysis. A sample of 67 URLs from dates outside of the period of interest (both before and after) were coded by three researchers, yielding a 90.1% alignment on average across the categories (min. 81.1%, max. 100%) and an average Krippendorf coefficient of 0.827 (min. 0.704 and max. 0.944). Next, an inductive analysis of the topics addressed by the content was conducted to identify any thematic patterns that might emerge.
The code scheme includes URL content analysis (e.g. watching an entire video and reading a posted text) and considers the channel and its constitutive elements. Furthermore, instead of discarding unavailable content, we mapped it because we believe that it holds an important interpretive value based on the analysis of the behavior of the news during contexts of crisis, as we will see below, related to source quality and credibility. The code scheme is as follows:
-
Content analysis:
Content availability on the coding date (Yes/No)
Is the link accompanied by text? (Yes/No)
Linked content format: Text, image, audiovisual, not available.
Type of content: Interview, news piece/information, testimony/statement (therefore a subjective view), artistic, not available.
Political nature: Does the content (URL plus text) mobilize the users? (Yes/No)
Topic (inductive qualitative approach): Social outbreak (estallido), immigration, politics, patriotism, protests, other.
-
2. Basic Metadata
Date the content was published online.
Title
Description or first paragraph
-
3. Analysis of the channel:
Name of the channel
Source: YouTube, Twitter, Facebook, Instagram, website, other, not available.
Type of Source: Personal profile, topic-specific channel/profile, public group, private group, institutional profile/channel, public media outlet, private media outlet, not available.
Regarding the type of source, these categorizations emerged from an initial qualitative review, distinguishing between more established media outlets and those that are more informal and the product of individual initiatives: a channel linked to a government-owned or publicly funded media outlet is considered a public media outlet; a channel linked to traditional commercial media, whether mainstream or alternative, is considered a private media outlet; and a channel linked to a formally existing public or private organization, beyond the digital realm, is considered an institutional media outlet. Meanwhile, a channel linked by its name to a natural person, identifiable by their given name or an alias, is considered a personal profile; a channel associated with a specific topic, which is made explicit by its name and/or its description, is considered a thematic channel; and a private group is defined as an online community to which external users have no access, such as a private Facebook group.
Results
First, we will offer a general descriptive analysis of the corpus. We will then conduct a more detailed analysis of the period of interest. The goal is to describe the group in general based on their communication dynamics and news behavior (RQ 3) and to have a basis for comparison with the period of interest, as we move on to analyze the information sources shared (RQ1) and assess their credibility and quality (RQ2).
General descriptive analysis
The complete corpus consists of approximately 50,000 messages sent by 188 users over the course of five months with a peak of 1,173 messages on October 20 (graph 1). This shows what we expected to find: a significant relative increase in the group’s activities in the days leading up to and following the estallido.
The period of interest for this article produced 6,593 messages, 483 (7.3%) of which included a link shared with the group. Those links were duly coded. During the period of interest, the peak period was from the early morning hours of October 17 until the following day, between 1 and 2 a.m., when 195 messages were sent — more than three per minute — during that hour, six of which contained a URL (graph 2). The URL peak during the period was between 10 and 11 p.m. on October 19, when 14 URLs were shared. This may indicate a gap in the availability of content that could be used to discuss, digest and describe the event: first talk, then share sources.
Audit of URLs
URLs during the entire period
An analysis of the URLs shared by the users during the entire period also shows (graph 3) the patterns of URL publication (left) and communication (right) of each user as they embrace the functionalities and publish the different types of content allowed by the application: audio, documents, images, texts, URLs and videos. Of the 4,012 URLs shared during the entire period (5 months), 17% were shared by the user TEL007, who posted nearly twice as many URLs as the second most prolific user. TEL003 shared the most URLs relative to total content shared. One of every three messages posted by this user was a URL. Different users are more active regarding different activities or sources, an observation that prompts the need for further qualitative research in the future.
Top 20 users who published URLs most frequently (left) and all the types of content shared (right) — full period
While YouTube was the most shared platform throughout the period in which the data were captured, on 18O there was a clear increase in YouTube content, which represent more than 66% of the URLs shared that day compared to around 45% over the entire period (graph 4). YouTube is followed by Facebook then Twitter, but both to a much lesser extent. Other noticeable results are the secondary role of news sites and the appearance of BitChute, a low-moderation alternative platform to YouTube. BitChute’s content is characterized as “conservative news, politics, and conspiracy theories” (Trujillo et al., 2020, p. 140), many of whose authors are deplatformed YouTube users (Rauchfleisch; Kaiser, 2021). During the coding process, some digital forensics applied to unavailable content (e.g. searching for the unique YouTube string on search engines) led to extremist content on BitChute, confirming this sort of pipeline from YouTube to BitChute, which is a much less scrutinized platform.
Top five platforms shared by users comparative total (above), 18-o (middle) and after the estallido (below)
Graph 5 shows the list of sources by type: media, social media, YouTube, and other. Only 4% of the URLs shared are from media outlets, even indiscriminately counting official well-established Chilean press outlets like La Tercera and EMOL along with more emergent media such as El Desconcierto or El Líbero. Graph 5 displays users’ different news consumption habits. Some (TEL073, TEL007, TEL008) share more YouTube content, while others connect the group to other social media content (which eventually also brings media content to the group), particularly Twitter and Facebook (Users TEL002 and TEL003 stand out with over 90% of their URLs from those sources.) Of those who most frequently share media sources, relatively speaking, only two users share more than 20% of their URLs based on the media (TEL019 and TEL032).
URLs during the estallido
We saw YouTube’s dominance on the day of the social uprising itself. This implies more than an inclination to share audiovisual content during the crisis. It suggests that there can be a multiplicity of sources — with different quality levels — given the diversity of YouTube channels. One important consideration is the number of sources that are not available less than three years after their publication: 40%.
The lack of availability of sources is an understudied phenomenon that has important implications (Alves, 2025), particularly regarding the reliability of sources shared in contexts of crisis in groups like the one studied here. The reason for their erasure is also important: some are removed by the user or platform, and some are unavailable because the privacy of the account increased. Half of the YouTube videos, one third of the Facebook content, 6% of tweets and websites and a few posts on Instagram were unavailable. The platform that provides the most information about the type of lack of availability is YouTube. More than half of the URLs (52%) were made private,9 30% are no longer available, 12% were removed because the users’ accounts were canceled, and 6% were related to a violation of the platform’s terms (see table 1).
An analysis of the types of content most frequently shared via URL highlights the relevance of testimonial or declarative content. This type of content presents a subjective vision, which we categorize in two variations: (1) the author(s) may share their opinion in a declarative fashion, or (2) the content may literally indicate their perspective, for example, by presenting content captured on their cell phone. In both cases, the quality or credibility of the content is not a decisive aspect for its sharing, based on the criteria reviewed above. Though the type of content is not necessarily associated with its quality, for there are both testimonial and declarative content that may be well informed, a closer look at the actual sources, as we coded the data, shows no such thing. The declarative videos are from ill-informed individuals via individual or thematic channels, rather than, for instance, someone with authority to interpret the current events. The testimonial videos offer glimpses of reality, but our analysis of this set of URLs shows that they are accompanied by poor-quality commentary from users who post or repost them. Furthermore, as the crisis intensified, the role of strictly news-based content decreased. While it was around 10 to 20% in the days leading up to the estallido, it dropped to less than 6% on 10/19/2019, the day after the protests peaked, when it could be assumed that people were trying to understand what was happening.
Graph 7 depicts changes in the topics addressed by the URLs shared during the period of interest. A shift is seen in the main topics of interest as the key date of the estallido, 18/10, approaches. At first, migration was the key issue, then it became the estallido social itself, which was referenced in 77% of the URLs on October 20. It is interesting to note that the term ‘Left’ also became very important during that period, since the protests were initially associated with this political alignment. Figure 1 shows the distribution of subtopics within the discussions related to the social outbreak, with terms related to the current situation, and many related to conspiracy theories, which we now know have been debunked, such as interference by foreign nationals such as pro-Chavez or Cuban groups. Other important sub-topics include distrust of the press and patriotic activism, the latter of which suggests that these groups can also be useful for mobilizing their members during situations like the one experienced at that moment in Chile.
As the word cloud below (figure 2) shows, the titles of the sources point to the main news topics: Institutions such as the leftist political coalition Frente Amplio, the UN (ONU), the Communist Party (Partido Comunista), the government (gobierno), the military (militares), the police (policia); key Chilean or regional political figures such as President Piñera, President Maduro, former dictator Pinochet, First Lady Cecilia Morel, Donald Trump, alt-right spokeswoman Teresa Marinovic, and members of Congress; attempts to describe or define the events, such as violence (violencia), war (guerra), chaos (caos), crisis, protests (protestas), social (social), awaken (despierta); locations such as Chile, Santiago, square (plaza), metro, downtown (centro); and solutions, such as impeachment (destitución), voting (votar), curfew (toque de queda), detained (detenido).
There are important individual differences in the availability of the content shared by each user. When we tracked the content posted by the ten users who shared the most URLs during the period analyzed, only one user’s shared content has over 80% of their content still available. While they are all on YouTube, 30% are media channels on the platform. Websites in general seem to be more stable than social media, with an 83% rate of content that was still available online when the analysis was conducted. Facebook is on the other extreme, with nearly 60% of all its content unavailable.
Discussion
Credibility of shared sources
We observed that as the estallido occurred, community members paid much more attention to YouTube than to other media outlets or platforms. During the period of interest, despite the presence of traditional media on YouTube, topic-specific channels and personal profiles, both ran by individuals with no journalistic background, received the most attention (46% and 23%, respectively). This may pose a risk to adequately informed deliberation about the events. Few YouTube links are from institutional channels (10%) or media channels (17%).
The literature suggests that credibility is associated with a series of variables related to the quality of information — which, as mentioned below, is not met to a great extent — and with the expertise and reliability of the source that disseminates information (Flanagin; Metzger, 2000). In this analysis, questions emerge about the credibility of the content shared and its sources. This is due to the simple fact that the sources are no longer available. Most of them are no longer found less than three years after the events10 and information sources that cease to be public for one reason or another lose credibility and the opportunity to be consulted or cited. Moreover, the fact that the information is not publicly accessible means that the sources and their content cannot be properly examined through, for example, data verification tasks or analyses like this one.
Another political symptom may be inferred by the videos that are unavailable on YouTube (see table 1). The fact that most (52%) videos were made private (that is, only certain accounts have access) raises questions about the reasons to do so. Videos that are unavailable (30%) should be those taken down by the people who posted them. Furthermore, some videos violate the terms of the platform (6%), and others belong to accounts that have been suspended (12%), which may point to repeated violations of the terms, such as the use of hate speech or propagating disinformation. This radiography entails issues of memory and accountability of the content and of the authors of the content. Alves’ (2025) analysis of the case of an attempted coup d’état in Brazil on January 8, 2022, detected channels that supported the coup which deleted 90% of their content, raising problems related to misinformation creation and diffusion as well as democratic integrity, and accountability of the authors. This legal vacuum has been used intentionally by politicians like former President Jair Bolsonaro. After he lost the elections, he accused the electoral authority of fraud and three hours later deleted the post (Vasconcelos, 2023).
Quality of shared sources
Regarding the quality of the sources shared, very few pass the basic filters suggested in the literature. Our examination of the full content behind the URLs shows an abundance of second-hand, decontextualized information without diverse or balanced perspectives or concern for formal communication aspects (Alessandri et al., 2001; De Pablos; Mateos, 2004; Gutiérrez-Coba; Salgado-Cardona; Gómez-Díaz, 2012). The lack of availability of many links also points to the low quality of host channels (Oliván; Ullate, 1999).
YouTube as the dominant source of content is problematic in other senses beyond that of low quality or low credibility channels. The literature shows that YouTube is ideal for the dissemination of radical content in audiovisual format for extreme right-wing groups, for example (Bratten, 2005; Cammaerts, 2009; Rodríguez Serrano; García Catalán; Martín Núñez, 2019). Their channels function as a repository — that is, they are used to launch content on other media, such as WhatsApp — and as social media channels because they can enlist subscribers who can interact with ‘likes’ and comments on videos. Minority groups can use YouTube to deploy their ideologies without the need for their messages to pass the filters of traditional media organizations. This is fertile ground for ‘partisan reasoners’ with high leadership goals (confirmation bias) and weak precision (source quality) (Lodge; Taber, 2000). Furthermore, although there is no consensus (Ledwich; Zaitsev, 2020), various studies point to a radicalizing role of the YouTube algorithm, which recommends more politically extreme content as the user browses platform content, promoting, in absence of regulation on said platforms (Bryant, 2020; Nicas, 2018; Tufekci, 2018), the ‘laissez faire’ of the dawn of the Internet (Corbeil; Rohozinski, 2021, p. 164). Over the course of this research project, the authors were also exposed to extreme content as they engaged in the content analysis of the URLs.
Soliloquies of Political (Mis)Understanding
Various communication patterns can be observed depending on the user: some shared many URLs, others focused on a single type of source such as YouTube or social media or more on a type of content such as audio, when led to different types of content leadership in the community. These descriptive observations are important due to the lack of research on MIM content, and serve as a point of comparison to future work. Nevertheless, perhaps the most interesting and unique finding of this study in the context of conversations in MIM is what we call soliloquies of political (mis)understanding.
Various studies have shown that simple exposure to the news is not sufficient to increase knowledge (Chaffee; Schleuder, 1986) and can even be prejudicial to knowledge of political affairs if news consumption is not active (Valenzuela; Bachmann; Bargsted, 2021). The perception that news will find an individual on social media, relieving that individual of the need to actively seek out the news (the News-Finds-Me perception) can be prejudicial to political knowledge (Gil de Zúñiga; Weeks; Ardèvol-Abreu, 2017; Lee, 2021) and even undermine political interest (Gil de Zúñiga; Diehl, 2019).
The fact that 55% of the links shared are not accompanied by text could be an empirical indicator that there is a lack of care or effort when it comes to involving the community in contextualizing the content. In the case of Chile’s social outbreak, it could be that the topic was implicit because there is very little content about anything other than the estallido itself. But the data demonstrate the opposite: there is more information about context during the protests, perhaps more effort to engage in conversation, while prior to the uprising, when there was a greater diversity of topics (see figure 1), behavior was less careful, with less text accompanying the links (graph 8).
In addition, one in 10 sources shared during the period was shared more than once. This suggests not only a convergence of sources among the users — a characteristic that could be identified with both filter bubbles and echo chambers (as operationalized by Bruns (2017) — but is also empirical evidence that shared content does not necessarily involve content read or learnt by other members of the community. While this may be expected, the data help magnify the phenomenon and raise some interesting future research questions.
Conclusions
This article focuses on the use of a group chat with a significant ideological bias to obtain information during an extraordinary event, namely a recent social outbreak in Chile, known as the estallido social. The opportunity to dive into a politicized group chat during times of crisis is of value to the field. The descriptive data sets some communicative and sharing patterns that could be used comparatively as more research is done within this type of restrictive environment.
Digging a little deeper, in the case considered in this article, the data suggest implicit recognition of sources as endowed with expertise and reliability even when they very explicitly fail to meet the most basic journalistic parameters. Their low quality and credibility are not the only important factors. In contexts with high levels of homophily like a WhatsApp group focused on political-ideological issues (in this case, the ‘patriots’), exposure to media with marked political bias possibly leads to polarization through the constant confirmation of previous beliefs. The opposite is also true: polarization possibly leads to increased selective exposure to like-minded media (Stroud, 2010). In the context of discussions of an important social crisis in digital environments, or even worse, enclosed highly political messaging groups, the higher need for orientation is reflected not only in the detection of relevance of information about the crisis and the following pursuit to lower levels of uncertainty, but also in a forgotten component, taken for granted in previous generations of communication research: the expectation that reliable sources are likely to be available. In high-choice digital information environments, the difficulty to distinguish between levels of source credibility and reliability, especially when not under public scrutiny — such as MIM groups — should be further inquired.
This study was conducted in the context of debates around technical and ethical dilemmas related to the study of MIM platforms (Barbosa; Milan, 2019). This in no way means that the only unit of analysis of interest is the shared URLs. Nor does it mean that the study of shared URLs is exempt from limitations: the fact that a URL is shared does not imply it was opened by one, two or all users. They may very well have zero-rating data restrictions for WhatsApp, which was almost a default in Chile at the time, while navigation on other platforms could imply the consumption of data from the user’s allowance. As the social importance of MIMs increases, new procedures could help comply with ethical standards while simultaneously advancing the research agenda of the social phenomena associated with their use. This type of study reveals two factors: (i) the difficulty and need to consider regulation of the use of digital platforms or other forms of advancement to produce a healthier, higher quality news ecosystem, particularly on the platforms that stand out in this article (WhatsApp, YouTube, Facebook and Twitter/X); and (ii) the demand for better digital skills beyond merely operational ones and those related to news and social interaction (Van Deursen; Helsper; Eynon, 2016), for example, how to find and assess quality information, and when, how and with whom to share it. News literacy should make users more prudent (Vraga; Tully, 2021).
This article raises many interesting questions that merit further exploration, as indicated throughout the text. However, the study is based on a single group with unique characteristics. Possible future objects or strategies we foresee include: comparative studies; studying various similar groups; one or more heterogeneous groups; studies of groups with a different orientation (the radical Left, for example); and various platforms.
References
- Alessandri, F.; Edwards, C.; Pellegrini, S.; Puente, S.; Rozas, E.; Saavedra, G.; Porath, W. VAP: um sistema métrico da qualidade jornalística. Cuadernos de la Información, [ s. l. ], v. 14, p. 112-120, 2001.
- Alves, M. Moderação de conteúdo no Youtube: análise da remoção de vídeos da eleição de 2022 ao 08 de janeiro de 2023. Intercom - Revista Brasileira de Ciências da Comunicação, São Paulo, v. 48, p. E2025113, 2025.
-
Amnesty International. Ojos sobre Chile: Violencia policial y responsabilidad de mando durante el Estallido Social. [ S. l. ]: Amnesty International, Oct. 2020. Available from: https://www.amnesty.org/es/wp-content/uploads/sites/4/2021/05/AMR2231332020SPANISH.pdf Accessed: 5 Aug. 2023.
» https://www.amnesty.org/es/wp-content/uploads/sites/4/2021/05/AMR2231332020SPANISH.pdf - Aruguete, N.; Calvo, E.; Ventura, T. News by popular demand: Ideological congruence, issue salience, and media reputation in news sharing. The International Journal of Press/Politics, [ s. l. ], v. 28, n. 3, p. 558-579, 2023.
- Barbosa, S.; Milan, S. Do not harm in private chat apps: Ethical issues for research on and with WhatsApp. Westminster Papers in Communication and Culture, [ s. l. ], v. 14, n. 1, 2019.
- Bauman, Z. Comunidad. Buenos Aires: Siglo Veintiuno, 2003.
- Bawden, D.; Robinson, L. Information Overload: An Introduction. In: Thompson, W. R. (ed.). Oxford Research Encyclopedia of Politics. Oxford: Oxford University Press, 2020.
- Bishop, L.; Gray, D. Ethical Challenges of Publishing and Sharing Social Media Research Data. In: Woodfield, K. (ed.). The Ethics of Online Research 2. Bingley: Emerald Publishing Limited, 2017. p. 159-187.
- Bratten, L. C. Online zealotry: La France du peuple virtuel. New Media & Society, [ s. l. ], v. 7, n. 4, p. 517-532, 2005.
-
Bruns, A. Echo chamber? What echo chamber? Reviewing the evidence. In: 6th Biennial Future of Journalism Conference (FOJ17), 14-15 Sept., Cardiff, United Kingdom. (Unpublished). Available from: https://eprints.qut.edu.au/113937/ Accessed: 5 Aug. 2023.
» https://eprints.qut.edu.au/113937/ - Bryant, L. V. The YouTube Algorithm and the Alt-Right Filter Bubble. Open Information Science, [ s. l. ], v. 4, p. 85-90, 2020.
- Cammaerts, B. Radical pluralism and free speech in online public spaces: The case of North Belgian extreme right discourses. International Journal of Cultural Studies, [ s. l. ], v. 12, n. 6, p. 555-575, 2009.
- Chaffee, S. H.; Schleuder, J. Measurement and effects of attention to media news. Human Communication Research, [ s. l. ], v. 13, n. 1, p. 76-107, 1986.
-
Chile: el Ejército levantó el toque de queda en Santiago y Valparaíso. Infobae, [ s. l. ], 26 out. 2019. Available from: https://www.infobae.com/america/america-latina/2019/10/26/chile-el-ejercito-levanto-el-toque-de-queda-en-santiago-y-valparaiso/ Accessed: 10 Nov. 2024.
» https://www.infobae.com/america/america-latina/2019/10/26/chile-el-ejercito-levanto-el-toque-de-queda-en-santiago-y-valparaiso/ - Corbeil, A.; Rohozinski, R. Managing Risk: Terrorism, Violent Extremism, and Anti-Democratic Tendencies in the Digital Space. In: Cornish, P. (ed.). The Oxford Handbook of Cyber Security. Oxford: OUP Oxford, 2021. p. 163.
- De Pablos, J. M. ; Mateos, C. Estrategias informativas para acceder a un periodismo de calidad, en prensa y TV. Ámbitos , [ s. l. ], v. 11-12, p. 341-365, 2004.
- Festinger, L. La teoría de la disonancia cognoscitiva (extracto por A. Ovejero). Psicothema, [ s. l. ], v. 5, n. 1, p. 201-206, 1993.
- Fidel, R. The case study method: A case study. Library and Information Science Research, [ s. l. ], v. 6, n. 3, 273-288, 1984.
- Fischle, M . Mass response to the Lewinsky scandal: Motivated reasoning or Bayesian updating? Political Psychology, [ s. l. ], v. 27, n. 1, p. 135-151, 2000.
- Flanagin, A. J. ; Metzger, M. J. Perceptions of Internet information credibility. Journalism and Mass Communication Quarterly, [ s. l. ], v. 77, p. 515-540, 2000.
- Franzke, A.; Bechmann, A.; Zimmer, M.; Ess, C. Internet Research: Ethical Guidelines 3.0. [S. l.]: Association of Internet Researchers, 2020.
- Fukuyama, F.; De Salcedo, D. P. Confianza. Buenos Aires: Atlántida, 1996.
-
Galarraga, N. El Tribunal Supremo de Brasil ordena bloquear Telegram para frenar la desinformación. El País, Madrid, 18 mar. 2022. Available from: https://elpais.com/internacional/2022-03-18/el-tribunal-supremo-de-brasil-ordena-bloquear-telegram-para-frenar-la-desinformacion.html Accessed: 10 Nov. 2023.
» https://elpais.com/internacional/2022-03-18/el-tribunal-supremo-de-brasil-ordena-bloquear-telegram-para-frenar-la-desinformacion.html - Gil de Zúñiga, H.; Ardèvol-Abreu, A.; Casero-Ripollés, A. WhatsApp political discussion, conventional participation and activism: exploring direct, indirect and generational effects. Information, Communication & Society, [ s. l. ], v. 24, n. 2, p. 201-218, 2021.
- Gil de Zúñiga, H.; Diehl, T. News finds me perception and democracy: Effects on political knowledge, political interest, and voting. New Media & Society, [ s. l. ], v. 21, n. 6, p. 1253-1271, 2019.
- Gil de Zúñiga, H. ; Weeks, B. ; Ardèvol-Abreu, A. Effects of the news-finds-me perception in communication: Social media use implications for news seeking and learning about politics. Journal of Computer-Mediated Communication , [ s. l. ], v. 22, n. 3, p. 105-123, 2017.
- Gutiérrez-Coba, L. M. ; Salgado-Cardona, A. ; Gómez-Díaz, J. A. Calidad vs. credibilidad en el periodismo por internet: batalla desigual. Observatorio, [ s. l. ], v. 6, n. 2, p. 157-176, 2012.
- Hall, R. The strategic analysis of intangible resources. Strategic Management Journal, [ s. l. ], v. 13, p. 135-144, 1992.
- Herrada Hidalgo, N.; Santos, M.; Barbosa, S. Affordances-driven ethics for research on mobile instant messaging: Notes from the Global South. Mobile Media & Communication, [ s. l. ], v. 12, n. 3, p. 475-498, 2024.
- Johns, A. 'This will be the WhatsApp election': Crypto-publics and digital citizenship in Malaysia's GE14 election. First Monday, [ s. l. ], v. 25, n. 12, 2020.
- Kalogeropoulos, A. Who shares news on mobile messaging applications, why and in what ways? A cross-national analysis. Mobile media & Communication, [ s. l ], v. 9, n. 2, 336-352, 2021.
-
Karasz, P. What Is Telegram, and why are Iran and Russia trying to ban it? New York Times, New York, 2 maio 2018. Available from: https://www.nytimes.com/2018/05/02/world/europe/telegram-iran-russia.html Accessed: 10 Nov. 2024.
» https://www.nytimes.com/2018/05/02/world/europe/telegram-iran-russia.html -
Kargar, S.; McManamen, K. Censorship and collateral damage: analyzing the Telegram ban in Iran. Berkman Klein Center, Research publication, [s. l.], n. 2018–4, September 2018. Available from: https://cyber.harvard.edu/publication/2018/censorship-and-collateral-damage Accessed: 10 Nov. 2024.
» https://cyber.harvard.edu/publication/2018/censorship-and-collateral-damage - Karnowski, V.; Leiner, D. J.; Kümpel, A. S.; Leonhard, L. Worth to Share? How Content Characteristics and Article Competitiveness Influence News Sharing on Social Network Sites. Journalism & Mass Communication Quarterly, [ s. l. ], v. 98, n. 1, p.59-82, 2020.
- Kopko, K.C.; Bryner, S. M.; Budziak, J.; Devine, C. J.; Nawara, S. P. In the eye of the beholder? Motivated reasoning in disputed elections. Political behavior, [ s. l. ], v. 33, n. 2, p. 271-290, 2011.
- Kozinets, R.; Patterson, A.; Ashman, R. Networks of desire: How technology increases our passion to consume. Journal of Consumer Research, [ s. l. ], v. 43, n. 5, p. 659-682, 2017.
- Ledwich, M.; Zaitsev, A. Algorithmic extremism: Examining YouTube's rabbit hole of radicalization. First Monday, [ s. l. ], v. 25, n. 3, 2020.
- Lee, S. Probing the mechanisms through which social media erodes political knowledge: The role of the news-finds-me perception. In: Shen, F. (ed.). Social Media News and Its Impact. London: Routledge, 2021. p. 51-73.
- Lévy, P. Inteligencia colectiva. Por una antropologi´a del ciberespacio. Washington, DC: Organizacio´n Panamericana de la Salud, Marzo 2004.
- Lewandowsky, S. ; Ecker, U. K. H. ; Cook, J. Beyond misinformation: Understanding and coping with the "post-truth" era. Journal of Applied Research in Memory and Cognition, [ s. l. ], v. 6, n. 4, p. 353-369, 2017.
- Lodge, M.; Taber, C. Three Steps Toward a Theory of Motivated Political Reasoning. In: Lupia, A. Jr.; McCubbins, M. D.; Popkin, S. L. (org.). Elements of Reason. New York: Cambridge University Press, 2000. p. 183-213.
- Lottridge, D.; Bentley, F. R. Let's hate together: How people share news in messaging, social, and public networks. In: CHI '18: Conference on Human Factors in Computing Systems. Montreal, Canadá, 21-26 abr. 2018.
- Luna, J. P. ; Toro, S. ; Valenzuela, S. Amplifying Counter-Public Spheres on Social Media: News Sharing of Alternative Versus Traditional Media After the 2019 Chilean Uprising. Social Media + Society, [ s. l. ], v. 8, n. 1, 2022.
- McCombs, M.; Weaver, D. Voters' Need for Orientation and Use of Mass Communication. In: International Communication Association annual convention, Montreal, p. 25-28, abr. 1973.
- Milan, S. ; Barbosa, S. Enter the WhatsApper: Reinventing digital activism at the time of chat apps. First Monday, [ s. l. ], v. 25, n. 12, 2020.
-
Newman, N.; Fletcher, R.; Robertson, C. T.; Eddy, K.; Nielsen, R. K. Digital News Report 2022. [ S. l. ]: Reuters Institute for the Study of Journalism, 2022. Available from: https://reutersinstitute.politics.ox.ac.uk/digital-news-report/2022 Accessed: 10 Nov. 2024.
» https://reutersinstitute.politics.ox.ac.uk/digital-news-report/2022 -
Newman, N.; Fletcher, R.; Schulz, A.; Andi, S.; Nielsen, R. K. Digital News Report 2020. [ S. l. ]: Reuters Institute for the Study of Journalism, 2020. Available from: https://www.digitalnewsreport.org/survey/2020/ Accessed: 10 Nov. 2024.
» https://www.digitalnewsreport.org/survey/2020/ -
Nicas, J. How YouTube Drives People to the Internet's Darkest Corners. The Wall Street Journal, [ s. l. ], 7 Feb. 2018. Available from: https://www.wsj.com/articles/how-youtube-drives-viewers-to-the-internets-darkest-corners-1518020478 Accessed: 10 Jul. 2023.
» https://www.wsj.com/articles/how-youtube-drives-viewers-to-the-internets-darkest-corners-1518020478 - Nir, L. Motivated reasoning and public opinion perception. Public Opinion Quarterly, [ s. l. ], v. 75, n. 3, p. 504-532, 2011.
- Oliván, J. A. S.; Ullate, J. M. A. Criterios para evaluar la calidad de las fuentes de información en Internet. Scire: representación y organización del conocimiento, [s. l.], v. 5, n. 2, p. 99-113, 1999.
- Ozawa, J. V.; Woolley, S. C.; Straubhaar, J.; Riedl, M. J.; Joseff, K.; Gursky, J. How Disinformation on WhatsApp Went From Campaign Weapon to Governmental Propaganda in Brazil. Social Media + Society, [ s. l. ], v. 9, n. 1, 2023.
-
Rauchfleisch, A.; Kaiser, J. Deplatforming the far-right: An analysis of YouTube and BitChute. SSRN, [ s. l. ], June 15 2021. Available from: http://dx.doi.org/10.2139/ssrn.3867818 Accessed: 10 Nov. 2024.
» http://dx.doi.org/10.2139/ssrn.3867818 - Rodríguez Serrano, A.; García Catalán, S.; Martín Núñez, M. Estrategias narrativas audiovisuales de desinformación en YouTube de la nueva extrema derecha europea. El profesional de la información, [ s. l. ], v. 28, n. 3, 2019.
- Rossini, P.; Stromer-Galley, J.; Baptista, E. A.; Veiga de Oliveira, V. Dysfunctional information sharing on WhatsApp and Facebook: The role of political talk, cross-cutting exposure and social corrections. New Media & Society, [ s. l. ], v. 23, n. 8, p. 2430-2451, 2021.
-
Salazar, G. El «reventón social» en Chile: Una mirada histórica. Nueva Sociedad, Buenos Aires, out. 2019. Available from: https://nuso.org/articulo/protestas-Chile-estudiantes-neoliberalismo/ Accessed: 20 Sep. 2024.
» https://nuso.org/articulo/protestas-Chile-estudiantes-neoliberalismo/ - Saldaña, M.; McGregor, S.; Johnson, T. Mind The Gap! The Role of Political Identity and Attitudes in the Emergence of Belief Gaps. International Journal of Public Opinion Research, [ s. l. ], v. 33, n. 3, p. 607-625, 2021.
- Salikov, A . Social media in Russian politics. Politologija, [ s. l. ], v. 99, n. 3, p. 64-92, 2020.
-
Santiago bajo estado de emergencia tras manifestaciones. Cooperativa, Santiago, Chile, 19 out. 2019. Available from: https://www.cooperativa.cl/noticias/pais/transportes/metro/santiago-bajo-estado-de-emergencia-tras-manifestaciones/2019-10-19/032438.html Accessed: 14 Aug. 2023.
» https://www.cooperativa.cl/noticias/pais/transportes/metro/santiago-bajo-estado-de-emergencia-tras-manifestaciones/2019-10-19/032438.html - Santos, J. G. B.; Freitas, M.; Aldé, A.; Santos, K.; & Cardozo Cunha, V. C. WhatsApp, política mobile e desinformação: a hidra nas eleições presidenciais de 2018. Comunicação & Sociedade, São Bernardo do Campo, v. 41, n. 2, p. 307-334, 2019.
- Santos, M.; Saldaña, M.; Tsyganova, K. Subversive affordances as a form of digital transnational activism: The case of Telegram's native proxy. New Media & Society, [ s. l. ], v. 26, n. 1, p. 131-153, 2021.
- Santos, M.; Valenzuela, S. Changing Media Landscapes and Political Participation. In: Grasso, M.; Giugni, M. (org.). The Oxford Handbook of Political Participation. United Kingdom: Oxford University Press, 2022. p. 841-857.
- Shirky, C. Here comes everybody: How change happens when people come together United Kingdom: Penguin, 2009.
-
Statista. Most popular social networks worldwide as of January 2022, ranked by number of monthly active users. Statista, [ s. l. ], 2022. Available from: https://web.archive.org/web/20220328184308/https://www.statista.com/statistics/272014/global-social-networks-ranked-by-number-of-users/ Accessed: 05 Mar. 2026.
» https://web.archive.org/web/20220328184308/https://www.statista.com/statistics/272014/global-social-networks-ranked-by-number-of-users/ - Stewart, N. K.; Al-Rawi, A.; Celestini, C.; Worku, N. Hate Influencers' Mediation of Hate on Telegram: "We Declare War Against the Anti-White System." Social Media + Society, [ s. l. ], v. 9, n. 2, 2023.
- Stroud, N. J. Polarization and partisan selective exposure. Journal of Communication, [ s. l. ], v. 60, n. 3, p. 556-576, 2010.
-
Subsecretaría de Telecomunicaciones - Subtel. Usuarios de Internet móvil consumen 10 GB en promedio al mes y número de teléfonos móviles sube a 26 millones. Subtel, Santiago, 18 Oct. 2019. Available from: https://www.subtel.gob.cl/usuarios-de-internet-movil-consumen-10-gb-en-promedio-al-mes-y-numero-de-telefonos-moviles-sube-a-26-millones/ Accessed: 5 Mar. 2026.
» https://www.subtel.gob.cl/usuarios-de-internet-movil-consumen-10-gb-en-promedio-al-mes-y-numero-de-telefonos-moviles-sube-a-26-millones/ - Tönnies, F. Comunidad y asociación. Granada: Comares, 2009.
- Tenenboim, O. ; Kligler-Vilenchik, N. The Meso News-Space: Engaging with the News between the Public and Private Domains. Digital Journalism, [ s. l. ], v. 8, n. 5, p. 576-585, 2020.
- Treré, E. The banality of WhatsApp: On the everyday politics of backstage activism in Mexico and Spain. First Monday, [ s. l. ], v. 25, n. 12, 2020.
- Trujillo, M.; Gruppi, M.; Buntain, C.; Horne, B. D. What is BitChute? Characterizing the "Free Speech" Alternative to YouTube. In: HT Conference - Hypertext and Hypermedia, 31. 2020, USA, Proceedings of the 31st ACM Conference on Hypertext and Social Media, July 2020. p. 139-140.
-
Tufekci, Z. The YouTube Algorithm and the Alt-Right Filter Bubble. New York Times, New York, 10 Mar. 2018. Available from: https://www.nytimes.com/2018/03/10/opinion/sunday/youtube-politics-radical.html Accessed: 10 Nov. 2024.
» https://www.nytimes.com/2018/03/10/opinion/sunday/youtube-politics-radical.html -
Valenzuela, S.; Bachmann, I.; Bargsted, M. The personal is the political? What do WhatsApp users share and how it matters for news knowledge, polarization, and participation in Chile. Digital Journalism, [ s. l. ], v. 9 n. 2, p. 155-175, 2021. Available from: https://doi.org/10.1080/21670811.2019.1693904 Accessed: 10 Jul. 2023.
» https://doi.org/10.1080/21670811.2019.1693904 - Valenzuela, S. ; Halpern, D. ; Katz, J. E. ; Miranda, J. P. The Paradox of Participation Versus Misinformation: Social Media, Political Engagement, and the Spread of Misinformation. Digital Journalism, [ s. l. ], v. 7, n. 6, p. 802-823, 2019.
- Valeriani, A.; Vaccari, C. Political talk on mobile instant messaging services: A comparative analysis of Germany, Italy, and the UK. Information, Communication & Society, [ s. l. ], v. 21, n. 11, p. 1715-1731, 2017.
- Van Aelst, P.; Strömbäck, J.; Aalberg, T.; Esser, F.; De Vreese, C.; Matthes, J.; Hopmann, D.; Salgado, S.; Hubé, N.; Stepinska, A.; Papathanassopoulos, S.; Berganza, R.; Legnante, G.; Reinemann, C.; Sheafer, T.; Stanyer, J. Political Communication in a High-Choice Media Environment: A Challenge for Democracy? Annals of the International Communication Association, [ s. l. ], v. 41, n. 1, p. 3-27, 2017.
- Van Deursen, A. J.; Helsper, E. J.; Eynon, R. Development and validation of the Internet Skills Scale (ISS). Information, Communication & Society, [ s. l. ], v. 19, n. 6, p. 804-823, 2016.
-
Vasconcelos, R. Bolsonaro posta e apaga fake news em que questiona vitória de Lula. CNN Brasil, [ s. l. ], 11 jan. 2023. Available from: https://www.cnnbrasil.com.br/politica/bolsonaro-divulga-video-que-questiona-vitoria-de-lula-mas-apaga-publicacao/ Accessed: 10 Dec. 2025.
» https://www.cnnbrasil.com.br/politica/bolsonaro-divulga-video-que-questiona-vitoria-de-lula-mas-apaga-publicacao/ - Vraga, E. K.; Tully, M. News literacy, social media behaviors, and skepticism toward information on social media. Information, Communication & Society, [ s. l. ], v. 24, n. 2, p. 150-166, 2021.
- Weaver, D. H. Audience need for orientation and media effects. Communication Research, [ s. l. ], v. 7, n. 3, p. 361-373, 1980.
-
Wolton, D. Dominique Wolton: "O elo indispensável, o receptor, é também o mais frágil". [Interview given to] NPC - Núcleo Piratininga de Comunicação. Revista do Fórum Nacional de Democratização da Comunicação, Rio de Janeiro, 8 jul. 2010. Available from: http://nucleopiratininga.org.br/dominique-wolton-o-elo-indispensavel-o-receptor-e-tambem-o-mais-fragil/ Accessed: 10 Nov. 2024.
» http://nucleopiratininga.org.br/dominique-wolton-o-elo-indispensavel-o-receptor-e-tambem-o-mais-fragil/ - Yin, R. Case study research and applications. Design and Methods (Sixth Edition). Los Angeles: SAGE, 2018.
- Zuckerman, E. New media, new civics? Policy & Internet, [ s. l. ], v. 6, n. 2, p. 151-168, 2014.
-
1
This is the first time the dataset is being used. It is available upon reasonable request made directly to the authors, to protect the privacy of the users implied.
-
3
A preliminary version of the results was presented at the preconference “Media & Communication in Global Latinidades” prior to the 73rd ICA Annual Conference on May 1, 2023. The authors report they have no competing interests to declare.
-
7
After Donald Trump took office the Meta project that verified information was ended. Details here: https://www.facebook.com/formedia/mjp/programs/third-party-fact-checking.
-
8
We will not disclose the precise name of the group to preserve its members’ anonymity and the safety of the researchers.
-
9
At the time of the research, YouTube allowed users to set a video’s visibility in three ways: (i) public, for anyone, indexed in the search; (ii) not listed, for anyone who has the link; and (iii) private, only visible to those who log in using an email registered by the channel owner. In this case, it is the private option, as we were unable to access the content even using the link. (Source: https://support.google.com/youtube/answer/157177)
-
10
The authors offer a “mea culpa” for not finishing the coding process faster, considering known issues around content persistence on social media. Discussions about how to deal with this kind of data from both methodological and ethical perspectives — to ensure significant results and not harm anyone (either researchers or the individuals researched) — and the coding process took more time than usual.
-
2
The first author received funding from the Agencia Nacional de Investigación y Desarrollo (ANID) through the Fondo Nacional de Desarrollo Científico y Tecnológico, grant Nº 11230980 and through the Fondo de Financiamiento de Centros de Investigación en Áreas Prioritarias, grant Nº NCS2024_007.
-
Associate Editor:
Fabíola Brigante Del Porto https://orcid.org/0000-0002-0672-1859
-
Editor-in-Chief:
Rachel Meneguello https://orcid.org/0000-0001-8288-0280
This is the first time the dataset is being used. It is available upon reasonable request made directly to the authors, to protect the privacy of the users implied.
Source: Created by the authors with original data.
Source: Created by the authors with original data.
Source: Created by the authors with original data.
Source: Created by the authors with original data.
Source: Created by the authors with original data.
Source: Created by the authors with original data.
Source: Created by the authors with original data.
Source: Created by the authors with original data.
Source: Created by the authors with original data.
Source: Created by the authors with original data.