CONNECTIVE ACTION AND CONSERVATIVE MOBILIZATION: AN ANALYSIS OF DIGITAL ACTIVISM AGAINST THE SCHOOL CURRICULUM WITH GENDER APPROACH IN PERU

 

Raúl Castro-Pérez

Scientific University of the South. Peru.

ccastroperez @ cientifica.edu.pe 

 

Santiago Tejedor

Autonomous University of Barcelona. Spain.

santiago.tejedor@uab.cat

 

Luis M. Romero-Rodríguez

University of Sharjah. United Arab Emirates / ESAI Business School - Espíritu Santo University. Ecuador.

lromero@sharjah.ac.ae

 

Diego Páucar Villacorta

National University of San Marcos. Peru.

diegopaucarv@hotmail.com

 

This work was carried out within the framework of the doctoral program in Communication and Journalism at the Autonomous University of Barcelona. It was made possible thanks to the support of the Seed Fund for Teachers at the Scientific University of the South, Lima, Peru, with grant number 029-2023-PRO99.

How to cite this article / Standard reference:

Castro-Pérez, Raúl; Tejedor, Santiago; Romero-Rodríguez, Luis M., & Páucar Villacorta, D. (2026). Connective Action and Conservative Mobilization: An Analysis of Digital Activism Against the School Curriculum with Gender Approach in Peru. Revista Latina de Comunicación Social, 84, 1-33. https://www.doi.org/10.4185/RLCS-2026-2553

Date of Receipt: Sep. 29, 2025
Date of Acceptance: Oct. 17, 2025
Date of Publication: Dec. 18, 2025 

ABSTRACT

Introduction: This research critically examines the formative period of the conservative mobilization that arose in Peru against the gender-focused school curriculum, in line with the resurgence of similar mobilizations worldwide. Using the concept of connective action (Bennett & Segerberg, 2012) as a framework, this study evaluates the Facebook activity of emerging activist groups. Through their posts, these groups construct a conservative narrative that positions them as deliberative agents. Methodology: A database of 6,221 online posts was developed that triggered two large-scale marches in Lima, Peru, in 2017. The posts were then systematically analyzed using a machine learning model in a Jupyter Lab environment (Python 3.11). Results: The research demonstrates a significant correlation between message simplification, intensive audiovisual media use, and radicalized discourse, accompanied by substantial follower growth. Discussion: This article discusses how conservative groups in Peru organize digitally without hierarchical structures. The article argues that communication reflects and constitutes the organization through formats, symbolic nodes, and affective economy. Contrary to "collective inertia," episodic connections are observed to generate politically effective communities. Conclusions: Conservative mobilization in Peru constitutes a social movement that disseminates simplified and polarizing anti-gender discourses through connective action. This digital strategy shapes an exclusionary political identity, influences public and institutional debate, and poses challenges to social cohesion, democratic governance, and national education policies.

Keywords: Connective action, online politics, conservative mobilization, school curriculum, gender approach, gender ideology, affective polarization.

1. INTRODUCTION 

In 2017, after the Peruvian state implemented a new gender-focused school curriculum, various conservative groups—including religious organizations, citizen collectives, and self-described "pro-family" platforms—joined forces for a protest that quickly gained significant public visibility. This alliance sparked a mass mobilization in defense of the "natural family" and the "right to life," presenting their opposition to the educational policy as a demand for respect of parents' "fundamental rights" to decide their children's moral upbringing. Although this study is based on data collected during the initial phase of the mobilization, the discursive repertoires, networks of actors, and mobilization strategies developed then have remained relevant, escalating and becoming embedded in governmental and parliamentary policy. The conservative narrative's continuity is evident today in growing sociopolitical polarization, where the language of rights is used to oppose inclusion and diversity policies. This phenomenon is not isolated; it is part of a transnational trend in which conservative sectors challenge democratic common sense through digital platforms and emotional communication strategies. This article critically examines the genesis of this phenomenon, paying particular attention to Lima-based social actors who are articulating new forms of collective action outside the traditional party system. Due to their decentralized organizational structure and personalized digital activism, this phenomenon will be analyzed using the framework of "connective action" (Bennett & Segerberg, 2012). This concept refers to a modern form of mobilization that is facilitated by digital technologies and networked structures.

The concept of "connective action" has significantly reformulated traditional theoretical frameworks of social movements, offering a better understanding of their current forms of constitution and political organization. This notion emphasizes the importance of "personalized communication" as a mechanism for uniting individual interests, contrasting with previous organizational structures involving party membership, activism, and hierarchy. Digital platforms and environments have emerged as the main facilitators of these dynamics by providing the technological means to transform individual expressions into collective actions, which are episodic and focused on specific causes or issues.

However, the concept has not been without criticism and debate. Shahin and Ng (2022), for example, highlight its relevance while also pointing out its lack of precision in identifying differences and peculiarities among the wide range of possibilities that social movements have today. They argue that the individualized nature of interactions, the flexibility of networks, and the absence of permanent identity constructions impede the formation of sustainable organizations, instead establishing "collective inertias." Pond and Lewis (2019) point out that the concept of connective action inadequately weighs the different impacts of various interactive technologies and does not fully address the peculiarities of the cultural and ideological drivers that cause and characterize their actions.

Thus, the usefulness and weaknesses of the concept are applied in the analysis of conservative or right-wing movements. While some scholars argue that these movements have a greater capacity to use digital media for political communication than their ideological counterparts (Freelon et al., 2020; Yoshida et al., 2021; Schradie, 2019; Barnes, 2022), the role of digital platforms in enabling conservative political and social mobilization has received limited attention, particularly from the perspective of connective action (Blee, 2017). A key precedent for this research is Kasimov's (2023) work, in which the author analyzes a conservative, right-wing community and identifies the usefulness of connective action. This is due to its capacity to decentralize mobilizations, foster immediate action without hierarchies, maintain the anonymity of its practitioners, and generate a large space for extreme, experimental ideas and tactics driven by negative sentiment discourse.

Kasimov's work aligns with the mainstream of studies that find interactive media to be a formative space for generating and promoting extreme ideas as a process of constantly validating beliefs (Marwick et al., 2022). Along these lines, studies such as that of Lerman et al. (2024) demonstrate that interactions between users with similar ideologies tend to be positive. In contrast, interactions between users with opposing ideologies manifest not only disagreement, but also active aversion, anger, and ultimately, hatred. Iyengar et al. (2019) categorize this emotional gap as the new status quo of public discourse and establish the phenomenon of "affective polarization" as constant. This is the self-referential process of meaning-making that is framed not only in discussions based on topics or problems but also in a priori positions of contempt and distrust. The so-called "confirmation bias," which is characteristic of social media, reinforces radicalism (Ghani & Rahmat, 2023) and exacerbates internet-era phenomena such as "filter bubbles" and "echo chambers," which are well-documented in political communication literature (Ross-Arguedas et al., 2022; Pariser, 2011; Jamieson & Capella, 2010). In practice, demonizing the "other" becomes normalized, introducing stigmatizing adjectives and the demonization of opponents as tactical games in discursive exchanges (Romero-Rodríguez et al., 2023).

In recent years, conservative movements have adopted increasingly rigid and doctrinaire approaches, focusing primarily on "moral politics" (Castro-Pérez, 2022). From this traditionalist position, they have mobilized against sex education and access to abortion, among other issues, adopting inflexible defensive attitudes (McIvor, 2019). Thus, conservative movements and right-wing factions globally have organized themselves at world summits and international forums, competing with progressive global policies (Bob, 2012). In Latin America, anti-gender movements have greater public relevance due to their mass mobilizations in alliance with political elites and the media, thereby increasing their movement's visibility (Bárcenas Barajas, 2022). Within this context, their mobilizations have gained strength due to their ability to organize and politicize through digital media. Peru is an exemplary case of this, with conservative movements having an active presence in debates, strong activity on social platforms, and even in de facto politics, with representatives in the national parliament and various government bodies (Castro-Pérez, 2024; Duárez Mendoza, 2024). This research focuses on the initial stages of the mobilization.

2. OBJECTIVES

This research analyzes how various emerging political groups in Lima, Peru's capital, articulate conservative political mobilization through shared opposition to the gender-focused school curriculum. Specifically, the research examines how these groups construct a narrative that allows them to cohere as a network and consolidate a base of followers. It also explores how these groups position themselves as deliberative agents in the public sphere, influencing key debates on educational policies, gender identity, and sexual and reproductive rights.

Additionally, the research seeks to understand how extremist discourse and increasing polarization impact the configuration of the contemporary sociopolitical landscape. The research analyzes the resulting fragmentation of public debate and reinforcement of ideological divides around issues such as gender, education, and civil rights.

To this end, the literature review presents a series of studies on conservative political mobilization, anti-gender discourse, and polarization in digital environments, with cases collected worldwide and across various platforms. This literature review considers key concepts and themes that current academic literature discusses regarding the conservative phenomenon and will be useful in this regard. Later, in the results section, a selection of data and an analysis of the digital behavior of relevant actors in these groups on Facebook will be presented through a computational analytical model capable of processing large volumes of information generated by their posts. The objective is to identify discursive patterns, levels of interaction, and political positioning strategies in the digital environment.

A critical review of the literature reveals two limitations to consider: First, much of the research in Latin America is exploratory without progressing toward comparative or generalizable theoretical frameworks that transcend specific cases. Second, although there is precedent for using tools such as social network analysis and big data techniques to study social movements and political polarization (Acosta, 2019), these digital methodologies are rarely incorporated systematically in the region.

Considering these gaps, this study aims to answer the following research questions (RQ):

• RQ1: What role do connective actions, such as participating in public events or sharing online slogans, play in enabling and activating the social and political mobilization of contemporary conservatism in Peru?
• RQ2: How do the extreme discourses and polarization resulting from this mobilization affect the contemporary sociopolitical landscape, and how do they relate to growth trends in conservative political mobilization in Peru??

The following section provides a brief overview of the existing research to establish the theoretical and conceptual framework for the proposed research.

2.1. Online Mobilization and Connective Actions

The state-of-the-art review of online conservative movements presented below considers literature published in academic journals from Donald Trump's first inauguration in 2016 to the present (Rohlinger & Bunnage, 2017). These strategies foster affective polarization, conflict, and social mobilization by leveraging decentralization, virality, and algorithmic resources (Ross-Arguedas et al., 2022). Several studies show that, from a comparative perspective, the "architecture" of each platform shapes the communication repertoires of the far right (Kakavand, 2023). In environments such as Facebook, Twitter/X, and YouTube, replicability, connectivity, and scalability facilitate the mass dissemination of content, which is reinforced by the algorithmic amplification of controversial messages (Freelon et al., 2020; Żuk & Żuk, 2020). In contrast, alternative, encrypted platforms like Telegram and 4chan rely on the constant creation of channels after blocks or controls and the anonymity of administrators to ensure the network's continuity in the face of regulation (Koltai, 2020; Kasimov, 2023; Cuevas-Calderón et al., 2022).

These platforms' messaging strategies are characterized by reappropriating topics and hashtags to delegitimize progressive counter-audiences, a practice known as frame hijacking. Examples include the #120db campaign, which changed the meaning of #MeToo to include xenophobic demands (Knüpfer et al., 2022), and the use of #AllLivesMatter, which diverted attention from #BlackLivesMatter (Gallagher et al., 2018; Xu, 2020; Knüpfer et al., 2022). They also employ polarized narratives and disinformation (McNerney et al., 2022; Fejős et al., 2021; Benkler et al., 2018; Recuero et al., 2022). In health-related debates, some anti-vaccine influencers engage in strategic interactions with pro-vaccine accounts to discredit them and project an image of legitimacy simultaneously. This dynamic is reinforced by the use of highly active bots, which artificially amplify their visibility and reach on social media (Featherstone et al., 2020; Yuan et al., 2019).

These communities are not solely sustained by hatred, but rather by mechanisms known as "meso" (Collins, 2024). These mechanisms are activated in spaces where users construct a collective identity that is intermediate between the individual (micro) and the global (macro). Through narratives of belonging, in-group/out-group comparisons, and shared symbols, these mechanisms transform the "I" into a cohesive and superior "we." On the social network Gab, conspiracies and symbols reinforce this sense of belonging and legitimize groups as "truth seekers," though this comes with the risk of violence (Collins, 2024). Emotional mobilization occurs through sensorially provocative messages that intensify visceral reactions and reinforce group identity (Bakardjieva, 2023; Featherstone et al., 2020; Cuesta-Cambra et al., 2019). On the far-right forum Stormfront, political shocks, such as Obama's election in 2008, activated interaction rituals that transformed fear and shame into mobilizing outrage and a reinforced sense of "we-ness" (Törnberg & Törnberg, 2023). These findings are consistent with other studies that highlight anger and negativity as key drivers of conservative collective action online (Grover & Mark, 2019; Ferré-Pavia & Sambucetti, 2022).

Among elites and in political-media ecosystems, these dynamics manifest through "digital surrogate organizations," connecting distributed audiences with conservative parties (Bennett & Livingston, 2025). For example, QAnon evolved from a fringe conspiracy to a digital surrogate for the GOP through phases of growth, symbiosis with MAGA, and adoption of the election fraud narrative. This granted QAnon legitimacy among Republican elites (Lukito et al., 2025). Similarly, floating signifiers like "Critical Race Theory" act as semantic pings that synchronize hyperpartisan media with legislators, amplify interactions, and give visibility to more extreme factions linked to alt-tech platforms (Knüpfer et al., 2025).

Regarding platform tactics, these organizations exploit less regulated alternative spaces to disseminate radical content that could be banned on open and publicly accessible social networks (Koltai, 2020; Kasimov, 2023; Cuevas-Calderón et al., 2022). They also maintain a hostile relationship with traditional media and the most popular publicly accessible platforms, allowing them to circulate messages while evading or avoiding direct challenge (Freelon et al., 2020; Żuk & Żuk, 2020). Studies show that they employ combined amplification and mobilization tactics: from growth hacking techniques and grassroots strategies to expand their follower base (Davidson, 2021; Vico & Rey, 2020) to the use of bots for greater visibility (Yuan et al., 2019). Furthermore, they exploit the unique features of each platform. On Facebook, they replicate content through links to alternative media and use groups or pages as coordination hubs. On Twitter/X, they use combined hashtags and strategic controversy. On Instagram, they use personalized aesthetics and memes. On YouTube, they use React videos and automated recommendations based on watch time. On Telegram, they use decentralized content replication and administrator anonymity (Kakavand, 2023).

2.2. Discourses and Practices that Polarize

Research on conservative groups in Latin American digital media has focused on analyzing their discourses and communication strategies using content analysis and digital ethnographies. Findings indicate strategies including misrepresenting concepts (López, 2020; Páez González & Peña García, 2018), constructing common enemies (Carrera Walling, 2021; Brito, 2020; Muro Ampuero, 2022) via mocking rhetoric (Cuevas-Calderón et al., 2022), disseminating disinformation to promote agendas (Barnes, 2022), and using pastoral language in grassroots communications (Meneses, 2019). These findings align with those of authors in the global literature. These movements often emphasize discourses such as individual freedom and distrust of scientific and governmental institutions (Peredo Rodríguez, 2022; Legua Aranibar, 2021), as well as anti-gender narratives, including an unwavering defense of the traditional family, a strict distinction between sex and gender, and a strict father figure (Reguer-Petit & Morabito, 2017; Castro-Pérez, 2024). These movements have gained prominence alongside the rise of influencers (Kessler et al., 2021) and the development of a transnational network of conservative activists (Sgró-Ruata, 2021), amidst a climate of polarization regarding health policy, justice, human rights, and gender issues.

The connective action framework enables an understanding of how digital networks of conservative movements develop structures that can become institutionalized as active organizers of the political and social agenda. Communication itself constitutes organization, generating collective identities and an affective economy of indignation and belonging that pressures parties, communities, and collectives toward more rigid views (Bennett & Livingston, 2025; Knüpfer & Klinger, 2025). QAnon (Lukito et al., 2025) and the use of floating signifiers, such as "Critical Race Theory," which align with conservative political parties (Knüpfer et al., 2025), are recent cases that demonstrate the effectiveness of conservative movements operating between digital bases and institutional structures. However, these logics are also replicated beyond political parties in networks such as Gab and Stormfront, Latin American anti-gender communities, and transnational conspiracy forums. Cohesion in these groups arises from symbols, shared narratives, and emotional practices that sustain common belonging and agendas (Collins, 2024; Törnberg & Törnberg, 2023).

2.3. Coordinated Link Sharing Actions

The digital space has become the central arena for political debate (Giglietto et al., 2020a). However, it has transitioned from a platform for promises of citizen participation to a space with systematic disinformation practices driven by political agents and extremist ideological movements. In this context, political astroturfing is introduced as the simulation of organic mobilization despite centralized coordination (Chagas, 2022). Coordinated Link Sharing Behavior (CLSB), consisting of sharing the same links within seconds of each other, has become a key indicator for detecting these practices (Giglietto et al., 2020b). Several studies have shown that CLSBs are closely linked to conservative and extremist groups in polarized contexts (Graham et al., 2020; Ayers et al., 2021). These groups act in a coordinated manner, either supporting right-wing figures or opposing left-wing policies (Gruzd et al., 2022; Chomel et al., 2023; Yu, 2022; Keller et al., 2020). They tend to operate in closed communities (Kim & Kim, 2023; Figeac et al., 2020) and combine legitimate sources with false narratives (Ayers et al., 2021; Kim & Kim, 2023). They also consciously and systematically articulate national and international communication networks (Gruzd et al., 2022; Rim et al., 2020).

The existence of coordinated actions to share links requires research methodologies that can capture their complexity. Therefore, the literature review shows the use of tools such as CrowdTangle (until its closure by Meta in 2024) and NodeXL, which are integrated with computational analysis to identify patterns of coordinated link sharing (CLSB) on social networks. The process usually starts with collecting posts based on keywords. Then, community detection and temporal coordination algorithms are applied to allow real-time tracking of groups sharing links at short intervals. This strategy generates traceability of digital mobilizations and has been applied in the analysis of recent elections in Italy (Giglietto et al., 2023) and the United States (Minici et al., 2024), for example.

Furthermore, specialized literature has documented how users interact with political content on social networks. Generally, users are more likely to respond to visual stimuli, engage in contentious discussions through comments, and share posts or links when expressing support or participating in public conversations (Kim & Yang, 2017; Ham et al., 2019; Tenenboim, 2022). It has been noted that sharing is not equivalent to simply reacting or commenting. While reactions and comments are usually motivated by short-lived interests, sharing is considered a stronger indicator of support for a post (Gerbaudo et al., 2023; Doroshenko & Tu, 2023). In fact, those who share content—even without reading it thoroughly—tend to be the most engaged users with a clear militant orientation (Sundar et al., 2024; Kim, 2018). Therefore, administrators of political pages carefully curate the content they disseminate to align messages with their ideological vision and strengthen the trust of their followers. These followers represent the main symbolic and organizational capital of these movements (Thompson et al., 2020).

3. METHODOLOGY

This research will analyze the early political and social mobilization of conservative groups in Peru, and their related actions, against the new gender-focused school curriculum, which originated in 2017. It pays particular attention to the activities in Lima, the country's capital, which was practically the movement's center of operations. Although the movement spread throughout the country and even to neighboring countries like Colombia, the study will focus on analyzing more than 6,000 posts on Facebook, emphasizing the degree of commitment and radicalization of their viewpoints based on an intense process of content creation and sharing. The actions included digital activity as well as in-person mobilization in the city, through the occupation of public spaces and mass marches in streets, bridges, and plazas. The study period spans from November 2016, when the actions began to be planned, to May 2017, taking into account two periods of high activity and both online and offline practices by the groups. The two moments are:

• Mobilization and occupation of public spaces, from January 9-11, 2017, when the connective action of the groups is expressed by agreeing on the concentrations and display of posters and banners in key points of the city, on the one hand.
• The mobilization of March 4, 2017, in which nearly one million people marched through the main streets of the city.

In this vein, this article uses two data sources: i. Publications on freely accessible Facebook pages related to conservative political and social mobilization in opposition to the Peruvian gender-focused school curriculum (collected via Apify), and ii . A database of links shared in Peru on this topic, obtained with permission from Facebook's Social Science One platform during the period under scrutiny. These data will be analyzed using a machine learning processing technique within a Jupyter Lab environment, from Python, to determine its efficiencies in viralizing content, and to assess to what extent these efficiencies foster expressions of coordinated online behaviors, which will mark the radicalization of their positions and the growth of the number of their followers.

3.1. Population and Sample

Conservative movements tend to focus their activity on a single social network, with Facebook being one of the primary platforms. In this context, it is crucial for these groups to effectively connect their online and offline networks to maximize their mobilization capacity (Theocharis et al., 2023). This study analyzes the role of conservative Facebook pages based on a key indicator: the sharing of external URLs within the platform to evaluate this connection. This metric approximates the reach and communicative impact of these pages beyond the immediate digital environment. This technique has been used in previous studies, including those by Fraxanet et al. (2025), and is useful for tracking the circulation of content and its potential to connect with offline audiences.

To ensure the thematic relevance of the selected URLs, a filter was applied based on the presence of keywords in the text accompanying the link's initial public share. The keywords—“género”, “educación”, “familia”, “homosex”, “gay”, “LGTB”, “currículo” e “ideología” ("gender," "education," "family," "homosexual," "gay," "LGBT," "curriculum," and "ideology")—were defined based on a preliminary review of public discourse and the semantic framework used by conservative actors during the study period. Only URLs whose first public appearance contained at least one of these terms were included. Links without textual context, duplicate publications, and URLs that were inactive or inaccessible at the time of analysis were excluded. This procedure yielded a robust, contextualized sample of 1,038 URLs for subsequent thematic analysis.

Data access and downloading were performed using a notebook provided by Meta that was specially configured with SQL queries. The keywords “género”, “educación”, “familia”, “homosex”, “gay”, “LGTB”, “currículo” e “ideología” ("gender," "education," "family," "homosexual," "gay," "LGBT," "curriculum," and "ideology") were used as queries to identify relevant URLs within the analysis period (n = 1,038). Associated texts were automatically classified using the BERTopic model, while external web pages were categorized manually.

Since Facebook was the central space for the analyzed mobilization, the Apify platform's Facebook Post Scraper tool was used to collect 6,221 posts disseminated across ten key, intentionally selected pages (see Table 1). Apify is a widely used tool in recent studies on social networks (Abas et al., 2025; Luriaga et al., 2025). The selection of these ten pages was based on two main criteria: (1) their high volume of URL sharing during the analysis period, and (2) their visibility metrics, such as the number of followers, posting frequency, and interactions generated. Their role in systematically disseminating content aligned with conservative positions typical of the Peruvian context was also considered (Barajas Bárcenas, 2022; Tello, 2019). The database is available online for consultation and verification (Páucar Villacorta, 2025).

The selected pages were considered key due to their discursive centrality and amplification capacity during the period of greatest conflict intensity (2016–2017). This characterization was based on a preliminary analysis of a database of URLs shared on Facebook. This analysis identified accounts with the highest number of shared links related to the school curriculum and related topics. In addition to the volume of posts, visibility and influence metrics were evaluated, such as the number of followers and the average number of interactions per post (e.g., reactions, comments, and shares). The frequency of conservative narrative frameworks in their content was also evaluated. The repeated presence of these pages at pivotal moments in the mobilization cycle, such as marches, boycott campaigns, and media disputes, confirmed their role as pivotal nodes in the structure of conservative digital discourse in Peru.

Table 1. Sample Characteristics

Name	Type	Followers (2024)[1]	ID	Recent Publications	Originals
Bethel Radio Lima	Religious	167,465	bethelradio	1,841	0.74
Coordinadora Nacional Pro Familia	Conservative	24,914	CONAPFAM	1,085	0.03
Únete a la fuerza	Political	178,234	unetealafuerza2021	8,720	0.59
Con Mis Hijos No Te Metas	Conservative	245,245	ConMisHijosNoTeMetasOficial	7,190	0.84
Impacto Evangelístico Lima	Religious	810,456	impactoevangelistico	577	0.3
Marcha Por La Familia - Perú	Conservative	68,000	marchaporfamiliaperu	402	0.44
Mami FL	Conservative	47,700	@salvemosalafamilia	299	0.51
No permitamos la Ideología de Género en el PERÚ	Conservative	2,171	nopermitamoslaideologia	208	0.64
Padres en Acción Perú	Conservative	5,470	Padresenaccionp	129	0.23
FEC	Religious	13,398	somosfcpe	109	0.82

Source: Elaborated by the authors.

4. RESULTS

4.1. Descriptive Statistics: The Radicalization Process and Negative Opinion

The two marches that were studied took place on January 9 and March 4, 2017. The sample was therefore divided into five periods: before, during, and after the two marches, as well as the periods between them. The intermediate periods lasted between one and three months, while the periods corresponding to the marches occurred one week before and one week after them. The period with the highest number of posts and the most pronounced peak was during the first march. However, the period with the most content sharing was after the second march (see Table 2). The analysis of topics reveals that at the beginning of the first march—a continuous activity spanning several days of occupying streets, bridges, and plazas and adorning them with signs and banners—messages became ideologized, taking the form of slogans that were standardized, simplified, and widely shared throughout the country (Figure 1). Advertisements and posters with messages such as "God instituted the family, don't change it," "I defend my children and their education," and, in particular, the phrase "Gender ideology" crossed out with a circle and a line through it demonstrate the radicalization of a stance expressed through slogans conveying intransigent viewpoints. The majority of posts were concentrated on the Facebook pages of Bethel TV, a Peruvian cable channel owned by the World Missionary Movement, and ConMisHijosNoTeMetas (Don't Mess With My Kids), with the caveat that most of their content promoted rallies and proselytizing events. Furthermore, the highest number of posts and daily interactions occurred between events. However, each page demonstrated its own dynamic. For example, ConMisHijosNoTeMetas was particularly relevant during the marches because it began to stand out as a slogan in itself (Figure 1).

Table 2 . General Descriptives Broken Down by Period (Percentages of Each Period, Unless Otherwise Indicated.)

		Before the marches	1st March	Between marches	2nd March	After the marches
Time ( approximate days )		60	15	40	15	90
Publications (total)		1,697	687	1,354	489	1,994
Facebook Pages
	@CONAPFAM	.17	.09	.16	.19	.20
	@salvemosalafamilia	.07	.10	.03	.04	.03
	Bethelradio	.32	.17	.26	.20	.36
	ConMisHijosNoTeMetasOficial	.04	.41	.15	.19	.04
	impactoevangelistico	.10	.05	.10	.08	.10
	marchaporlafamiliaperu	.05	.03	.07	.09	.08
	nopermitamoslaideologia	.04	.02	.04	.06	.03
	Padresaccionpe			.06	.03	.01
	Somosfecpe	.02	.00	.01	.02	.03
	unetealafuerza2021	.19	.12	.12	.10	.13
Interaction metrics ( averages )
	Reactions	564.6	429.7	604.6	1,055.3	450.9
	Comments	38.0	33.5	42.2	96.1	38.0
	Shared	251.4	248.2	348.6	607.3	277.0
	Characters (text)	127.4	113.9	132.1	109.9	124.9
	Views (videos)	1,833.7	1,363.6	2,558.7	8,819.5	2,038.8
Shared Content
	External links (presence)	.49	.40	.53	.46	.50
Opinion
	Negative	.31	.26	.35	.36	.29
	Neutral	.40	.64	.46	.51	.51
	Positive	.29	.10	.19	.13	.20
Content
	Links	.47	.39	.48	.42	.47
	Images	.36	.54	.32	.39	.31
	Text only	.03	.01	.01	.01	.00
	Video	.14	.07	.19	.18	.22
	Text and links	.47	.39	.48	.42	.47

Source: Elaborated by the authors.

Figure 1. Absolute and Relative Distribution of Publications from Selected Pages and Topics During the Event per Day (Above) and per Week (Center and Below).

Source: Elaborated by the authors.

The topics and types of pages that were shared the most during the study period reveal that discursive conflict is the strongest driver of content sharing (Figure 2). The most widely shared posts were about conflicts such as: i. A verbal and physical argument between Phillip Butters, a leading figure in the conservative movement and television presenter, and a journalist from the country's most-watched news program. This quarrel was broadcast live during a conservative march. ii. A campaign to cancel a telecommunications company for airing inclusive advertising. iii. Celebrations that the same telecommunications company withdrew its advertising from the program of the TV presenter involved in the aforementioned argument. The peak of this trend is visible in the following graph.

Figure 2. Topics and Types of Pages Shared in the Facebook URL Database (N = 1038).

Source: Elaborated by the authors.

In general terms, almost every post included in this study had more than one reaction, comment, or share. For every ten reactions, the pages generated an estimated five shares and one comment (Table 3). When estimating the bivariate correlations between normalized numerical values (shares and reactions/comments, as well as between reactions and comments alone), the ρ values ranged between .6 and .75 ( ≤0.01). However, the values drop between .34 and .41 ( ≤0.01) when comparing comments made after the event versus those made during it. In addition, the pages Impactoevangelistico, ConMisHijosNoTeMetasOficial, unetealafuerza2021, and marchaporlafamiliaperu generated four to ten times more interactions than the others.

Table 3. Presence and Average Interaction per Facebook Page.

		Reactions	Comments	Shared Content
Presence		1.0	.8	.92
Overall average		560.51	42.98	308.37
	@CONAPFAM	105.38	5.22	57.09
	@salvemosalafamilia	124.90	3.35	52.23
	Bethelradio	255.44	7.88	95.25
	ConMisHijosNoTeMetasOficial	1,024.47	120.36	516.19
	Impactoevangelistico	2,428.71	69.62	664.24
	Marchaporlafamiliaperu	684.74	109.54	638.48
	Nopermitamoslaideologia	4.12	1.49	1.81
	Padresaccionpe	26.65	3.35	10.79
	Somosfecpe	45.94	6.80	44.14
	unetealafuerza2021	509.81	84.96	744.24

Source: Elaborated by the authors.

The posts were primarily text accompanied by links, with images comprising 45% of the posts, user-uploaded images comprising 38%, and videos comprising 16%. The average text length was between 74 characters (nopermitamoslaideologia) and 177 characters (impactoevangelistico). After analyzing the texts, the most prominent categories were content celebrating collective achievements (16%), invitations to conventions or conferences (11.4%), posts criticizing gender ideology (13.6%), and critiques of gender ideology supported by news from other parts of the world (10.56%). The most shared posts included a journalistic report on corruption involving the main shareholder of a leading progressive media outlet in Peru and a report claiming that over one million people participated in the March 4^th protest. The most shared posts also included videos of interviews with Christian Rosas, the leader and founder of ConMisHijosNoTeMetas ("Don't Mess With My Children"), on the country's most politically influential news channel's flagship program. A similar interview with a former prime minister with strong conservative affiliations also appeared. The analysis revealed that Facebook pages dedicated to conservative or political proselytizing had more negative than positive text content (36%-66%), while the opposite was true for pages with religious content (23%-42%).

4.2. Polarization and the Fixation of Pro-Family and Pro-Life Views

The results of the interaction analysis revealed that entrenched positions on gender-inclusive education, gender identity diversity, and reproductive policies became more pronounced during this period. Shared links primarily led to content from pro-family or pro-life proselytizing pages (17.14%), Evangelical or Catholic media outlets (13.05%), and traditional (5.14%) and alternative (4.15%) media outlets (see Table 4). These findings were validated by the Facebook URL database (n = 1,038), which revealed that 33% of the links shared by platform users during the same period led to conservative proselytizing pages (parejasreales.net and conapfam.net), evangelical religious pages (impactoevangelistico.net and bethelradio.fm), and Catholic pages (aciprensa.com), as well as right-wing media outlets (actuall.com, tiempo26.com, and elciudadanodeapie.blog). Additionally, 25.4% of the links were to traditional media outlets (rpp.pe, Peru21.pe, LaRepublica.pe, DiarioCorreo.pe), and 24% were user-generated links to YouTube.com.

Table 4. Percentages of Posts According to the Types of Link Destinations Shared in Posts From Conservative Websites.

	Alternative Media	Blogs / Opinion	Christian Media	Pro-family / Pro-life	Others	Traditional media	Government	Social networks
@CONAPFAM	.02	.01	.01	.84	.01	.03	.01	.01
@salvemosalafamilia	.01		.03	.33		.03	.02	.03
bethelradio			.18		.05			.03
ConMisHijosNoTeMetasOficial	.02	.01	.01	.02		.06	.01	.02
impactoevangelistico			.69			.01
marchaporlafamiliaperu	.16	.01	.01	.04		.24	.04	.02
nopermitamoslaideologia	.06	.02	.04	.08	.01	.04	.02	.07
Padresaccionpe	.15	.02	.18	.14		.14	.05	.07
somosfecpe			.12	.01				.05
unetealafuerza2021	.14	.01	.01			.12	.03	.07

Source: Elaborated by the authors.

Nearly half of the posts with some type of image featured text (42.29%) or people (50.47%) as part of the images. Half (49.97%) of these images were photographs, followed by collages or posts with multiple images (18.62%), posters (9.64%), elaborate designs (6.47%), memes (5.9%), and screenshots (5.85%) (Figure 3; and Table 5). Different pages used different types of images; for example, the bethelradio page showed a greater number of images with diagrams or illustrations and portraits, a set mainly made up of church advertising. Finally, although the classification of images by similarity yielded categories with a balanced distribution, some pages had a higher percentage of informal advertising with anti-gender ideology messages than the rest (marchaporlafamiliaperu, nopermitamoslaideologia, unetealafuerza2021), others published images with religious messages (bethelradio, @salvemosalafamilia), and still others published more video captures (somosfecpe, bethelradio).

Figure 3. Count of Posts with Images per Main Topic of Images Over Time (Weeks).

Source: Elaborated by the authors.

Table 5. Percentages of Posts Based on the Presence of Text and People and Main Image Types Displayed in Conservative Page Posts.

	People	Text	Collage	Design	Meme	Photograph	Portrait	Poster	Screenshot	Video scene
@CONAPFAM	.60	.71	.08	.03	.16	.43		.08	.02	.20
@salvemosalafamilia	.74	.95	.10		.01	.56	.02	.09	.01	.19
Bethelradio	.70	.84	.14	.08	.01	.30	.03	.07	.01	.36
ConMisHijosNoTeMetasOficial	.75	.80	.13	.03	.04	.58	.01	.06	.02	.10
Impactoevangelistico	.36	.90	.05	.02	.03	.45	.02	.06	.01	.37
Marchaporlafamiliaperu	.60	.90	.06	.01	.08	.35	.03	.04	.09	.32
Nopermitamoslaideologia	.77	.84	.36	.05	.09	.18	.01	.06	.08	.15
Padresaccionpe	.33	.67		.13		.25				.63
Somosfecpe	.73	.85	.12		.04	.20		.20	.01	.43
unetealafuerza2021	.61	.96	.15		.13	.14	.02	.02	.30	.25

Source: Elaborated by the authors.

The majority (63.4%) of the videos were published by religious pages. However, these pages did not have the highest average number of views per video (see Table 6). The two pages directly linked to the mobilization that had the highest average number of views were marchaporlafamiliaperu, with 107,695 views, and ConMisHijosNoTeMetasOficial, with 94,564 views. The only political proselytizing page studied was also included (unetealafuerza2021, with 87,284 views). The remaining pages had fewer than 17,000 (impactoevangelistico) and 3,000 views, respectively.

Pages were more likely to share posts containing images of people or text ( = .42, p≤.01), video scenes ( = .24, p≤.01), image collages ( = .2, p≤.01), or photographs ( = .17, p ≤.01), as well as posts with messages directed at the reader ( = .22, p≤.01) or "formal" advertisements ( = .19, p≤.01). Concurrently, pages were more likely to share posts with titles supporting a pro-life viewpoint ( = .22, p ≤ .01), particularly if the link contained information about other countries ( = .29, p ≤ .01) or if the text generally expressed negative opinions ( = .20, p ≤ .01).

Conversely, interaction metrics (including shares) were significantly related to anti-gender critique posts containing images with "informal" posts (η²=[.01,.03], p≤0.01) and posts from the pages marchaporlafamiliaperu (η²=[.03,.08], p≤0.01) and bethelradio (η²=[.01,.03], p≤0.01).

Table 6. Percentages of Posts According to the Topics of the Images Used on Conservative Pages.

	Messages	Photograph (various)	People with signs	Mass of people	Interiors	Close -ups	Formal advertising	Informal posters	Screenshot	Video capture
@CONAPFAM		.21	.12	.19			.21	.07	.07	.14
@savethefamily	.35	.09	.06	.07	.03	.11	.13	.02	.01	.13
Bethelradio	.2	.05	.04	.01	.18	.13	.19	.01	.01	.17
Don't Mess With My Children Official	.06	.12	.23	.12	.04	.09	.09	.07	.08	.1
Evangelistic Impact	.02	.27	.08	.01	.01	.19	.07	.04		.32
March for the Family of Peru	.01	.06	.03	.03	.08	.15	.04	.32	.09	.2
Let's not allow ideology	.02	.07	.09	.02	.01	.02	.03	.57	.06	.11
Padresaccionpe		.57	.14							.29
Somosfecpe	.07	.11	.05	.03	.05	.2	.05	.08		.36
jointheforce2021		.04	.03	.01	.03	.05	.01	.63	.05	.16

Source: Elaborated by the authors.

The movement's most prominent themes were stories about the school curriculum featuring public figures (16.8%), which were particularly prevalent during the second march. The movement also covered both marches extensively (15%). These themes were highlighted with links to traditional media outlets and activist YouTube accounts. However, the other two most important themes—fighting gender ideology within churches (12.5%) and religious proselytizing (10.1%)—were only presented through links to religious websites.

Figure 4. Mobile Count of Original and Shared Posts and Mobile Average of Shares (Black Line) on Facebook Pages.

Source: Elaborated by the authors.

4.3. Anti-gender Emphasis and Coordinated Actions to Share Links

Correlational analysis shows how analyzed variables relate to the influence of marches and their anti-gender slogans. Figure 5 shows that the number of shares of the studied pages' posts increased during the second march, indicating a growing capacity to mobilize. Similarly, the amount of original content—that is, content produced by the pages themselves—also increased, though it peaked during the first march.

Figure 5 shows the results of the classification and regression models. During the analyzed period, the pages underwent heterogeneous changes. For example, the bethelradio page, the most active in the sample, was flooded with posts unrelated to politics after the marches. However, this analysis demonstrates that regardless of the pages' size and influence, the main changes in content sharing throughout the analyzed period were in the types of links shared. Before the marches, sharing was more closely related to links to Christian media (.118), pro-family media (.106), traditional media (.0723), social media (.067), and alternative media (.047), as well as images with text (.045) and people (.047). After the marches, there was a convergence in accentuating polarization and reinforcing pro-life and anti-gender viewpoints.

Figure 5. Balanced Importance of Variables per Period According to the Machine Learning Model Used.

Source: Elaborated by the authors.

These metrics show that the connected actions studied increased their content creation volume and focused on reinforcing their radicalized viewpoint. They also adjusted their communication strategies by increasing links to pro-family media and using images that align with the prevailing narrative. The short-term effect of the marches on the analyzed pages' content is clear: proselytizing and establishing a position. After the marches, however, the pages returned to their usual sharing patterns. This demonstrates their coordinated intention.

4.4. Mobilization Tactics and Connective Resources

The followers of the pages studied also changed their behavior significantly during the period studied. During the first march, users shared videos (=.171) and posts with images of people (=.150) or with text (=.135) en masse. Although the sharing of messages with negative or critical content prevailed throughout the study period, only during this brief interval was there a higher number of posts with positive content, even as predictors of sharing (075, =.812). 

In the period between marches, the most shared posts mainly contained criticism of gender ideology and references to news about anti-gender mobilization in other parts of the world (=.21). This same variable underwent a sudden change during the second march (=.023). (Table 7)

The most shared posts during and after the second march were the most numerous. The inclusion of videos was one of the determining characteristics (=.103, =.176). The same was the case for the use of photographs (=.095, =.084) and original images (=.053, =.178) featuring people (=.093, =.204) or text (=.069, =.176). It was then that the most shared posts during the second march were characterized by negative descriptions (=.117) and criticism directed at gender ideology (=.129). Finally, both at the beginning (=.043) and at the end (=.062), users again shared invitations to religious conferences and conventions to the same extent, confirming the fact that the same user base was observed.

Throughout the study period, a common thread was that followers preferred to share eye-catching posts, such as videos and images featuring people. Before the marches, negative descriptions and criticisms of gender ideology predominated. During the first march, however, positive messages became more prevalent. Subsequently, between the marches and during the second march, criticism and negativity resurfaced. Toward the end of the period and afterward, the content primarily consisted of graphics.

Table 7. Percentages of Post According to Topics Found in the Texts of the Posts, Broken Down Per Group and Per Page.

		@CONAPFAM	@salvemosalafamilia	bethelradio	ConMisHijosNoTeMetasOficial	impactoevangelistico	marchaporlafamiliaperu	nopermitamoslaideologia	Padresaccionpe	somosfecpe	unetealafuerza2021
Conservative collective action		.09	.37	.14	.73	.05	.21	.10	.07	.11	.03
	Invitations to marches		.02	.07	.02		.02	.03		.02
	Celebrating the movement's achievements	.08	.35	.07	.72	.05	.19	.07	.07	.09	.03
Collective religious action		.01	.25	.36	.03	.32	.03		.03	.34	.01
	Invitation to family conventions and congresses		.24
	Invitation to religious conventions and congresses	.01	.01	.27		.31	.01			.15	.01
	Other religious or church-related events			.05					.03	.13
	Report natural disasters and ask for help			.04	.02	.01	.01			.06
Ideological proselytism		.48	.13	.17	.21	.20	.51	.67	.67	.38	.82
	Anti-gender or pro-life critique with international references	.38	.03	.01	.02	.07	.24	.16	.14	.02	.02
	Anti-gender criticisms	.24	.10	.13	.16	.13	.29	.53	.50	.30	.12
	Criticism of traditional media	.01	.01	.03	.01	.01	.07	.07	.02	.06	.15
	Anti-left or pro-right political criticism	.01	.01	.01	.03	.01	.12	.07	.08		.55
	Pro-life criticisms	.22			.01	.05	.03		.08	.01
Religious proselytism		.01		.25		.32		.02	.02	.11	. 01
	Communicating religious messages			.12		.02		.01	.01	.06
	Communicating religious messages through Christian media			.12		.06			.01	.06	.01
Non-proselytizing information		.02	.21	.06	.01	.03	.01		.04	.04	.07
	Health information	.02	.21	.01	.01	.02			.01	.02
	Psychological-family information			.04		.01			.03	.02	.07

Source: Elaborated by the authors.

5. DISCUSSION AND IMPLICATIONS OF THE RESULTS

The research findings make it clear that the mobilization of conservative groups in Peru against the gender-focused school curriculum resulted from a sustained process of articulation, ideological radicalization, and digital viralization. This process was driven by groups with extensive activity on online platforms. This phenomenon can be explained by the connective action framework proposed by Bennett and Segerberg (2012). This framework describes new forms of collective action enabled by digital technologies. In these forms, mobilization is organized through the personalized exchange of content containing political, legal, moral, and emotional discourses.

A qualitative analysis of over six thousand Facebook posts revealed that the mobilizations occurred in a decentralized manner, without clearly defined leadership. Groups opposing the gender-focused approach do not depend on traditional organizational structures; rather, they coalesce through publications that generate trends and chains of meaning. These messages are personalized and reinterpreted through comments, memes, images, and videos that reinforce a common narrative centered on defending the "natural family" and rejecting "gender ideology." This personalization enables individuals with different levels of commitment to find meaning in collective action without a structured ideology. Contrary to Shahin and Ng's concept of "collective inertia," the Peruvian experience of 2016 and 2017 shows that online connections, even if they are episodic and personal, can foster politicized communities with a practical impact. However, these communicative dynamics tend to simplify discursive diversity and lead to radicalization that denies the legitimacy of differing viewpoints.

The results reveal clear empirical patterns. Mobilization peaked during the marches when messages were condensed into brief, radical slogans. Bethel TV and ConMisHijosNoTeMetas were among the pages with the greatest reach and acted as redistribution hubs. The most shared content was not devotional messages but rather episodes of public confrontation or boycott campaigns. This demonstrates that discursive conflict is more mobilizing than conventional propaganda. Additionally, it was found that format governs virality; videos and images with text or people were the strongest predictors of sharing, with a consistent ratio of five shares for every ten reactions. This shows that the economics of virality depend more on form than argumentative density. Media events such as the 2017 marches increased connective activity during and after their occurrence. Among the accounts with the largest following were ConMisHijosNoTeMetas, MarchaPorLaFamiliaPerú, and Bethel TV. Posts featuring people were the most shared (42%), followed by videos, collages, and texts with a negative slant (30-20%).

Figure 2 shows that disseminating "shared successes"—such as complaints against companies with progressive advertising or appearances by conservative leaders like Christian Rosas, founder of ConMisHijosNoTeMetas ("Don't Mess With My Kids"), in the media—was a key element of this connective action. These milestones served as nodes of meaning that reinforced the sense of collective political efficacy, even in the absence of formal leadership. Public figures like Rosas and TV presenter Phillip Butters emerged as symbolic references that sustained the set of communicative activities. Analyzing the most shared content using sentiment analysis revealed a strong pattern of affective polarization, consistent with Bennett's findings regarding the central role of emotionality in new forms of digital mobilization (Table 2). Content that aroused indignation, fear, or enthusiasm—particularly content appealing to the figure of the "ideological enemy"—was more likely to be shared, fueling a closed loop of ideological feedback.

Theoretically, the results contribute to our understanding of connective action in three ways. First, communication constitutes organization since patterns of format, links, and reference nodes produce thematic alignment and internal discipline. Second, the affective economy, which alternates between negativity, celebrations, and conflicts, explains cycles of cohesion more than formal structures do. Third, connective action becomes institutionalized in this case as slogans and symbolic markers, such as ConMisHijosNoTeMetas (KeepYourHandsOffMyChildren), transcend the episodic and remain as reusable repertoires. The combination of emotion, format, and symbols ensures that the network mobilizes, learns, and reproduces its viralization routines.

From a theoretical and methodological standpoint, it is important to note that authors such as Giglietto (2020a, 2020b) and Keller et al. (2020) describe what is known as "astroturfing" and "coordinated actions to share links" in the literature. In the Peruvian case, these actions have reached a hybrid stage involving in-person activities in public spaces such as city squares, streets, and bridges. The use of artificial intelligence was essential for processing and analyzing large volumes of Facebook data. Implementing Python scripts and using platforms like Apify enabled the collection of public information, the identification of viralization patterns and semantic networks, and the analysis of interactions and sentiment models that reveal trends no longer subject to deliberation. Limitations such as sample selection and data repetition are offset by the databases' versatility and strategic value.

In line with previous research, the results of this study confirm the effectiveness of frame hijacking and hashtag dispute strategies, as described by Knüpfer et al. (2022). In the Peruvian case, however, these strategies manifest on Facebook through link curation routines and customized visual formats that reinforce the circulation of anti-gender frames via superimposed videos and portraits. Although opposition to traditional media is common in other conservative movements (Freelon et al., 2020), in Peru, it was combined with the legitimization of local political and religious media as reference points bordering on disinformation.

The observed patterns of amplification and emotional sustainability resemble those documented by Törnberg and Törnberg (2023) in Stormfront, where political "shocks" activate interaction rituals that transform negative emotions into collective energy. In our study, the first march of 2017 functioned as an equivalent shock; it produced a surge of positive, celebratory messages that later returned to criticism. This revealed more complex emotional sequences than the monolithic negativity noted by Grover and Mark (2019). Although Facebook has a more centralized architecture, collectives used short-latency and rapid reactivation strategies that replicate cross-platform porting logic. In line with Collins (2024), cohesion is explained by not only hate but also meso-mechanisms of familiarity, the use of symbols, and conspiratorial narratives that consolidate a sense of "digital family."

With this in mind, the study's results confirm that the conservative movement's communicative effectiveness was not limited to producing visual content or spreading brief slogans virally. Rather, it operated on a deeper logic: identity simplification. This dynamic has been documented in global anti-gender mobilizations, such as La Manif Pour Tous in France. The objective is not to engage in deliberative debate but to construct a clear antagonist and offer a symbol of belonging that simplifies social conflicts and facilitates broad audience adherence (Reguer-Petit & Morabito, 2017). In this sense, the ConMisHijosNoTeMetas (KeepYourHandsOffMyChildren) page functioned not only as a channel for coordination and content dissemination but also as an enduring and defining slogan. From the perspective of connective action (Bennett & Segerberg, 2012), this evolution illustrates how movements can institutionalize reusable symbolic repertoires—in this case, a name, slogan, and moral narrative—that allow for sustainability and discursive expansion beyond the originating event.

 5.1. Limitations of the Study

This study is based on digital behavior data collected between 2016 and 2024 from conservative groups mobilized against the gender-focused school curriculum. The analysis, however, focuses specifically on the period between November 2016 and May 2017, which corresponds to the movement's founding stage. The study's main methodological limitation is the inability to access year-disaggregated metrics through Meta (Facebook's parent company), which hinders precise temporal interpretation of the phenomenon. Without statistics specific to the 2016–2017 period, the contextual analysis's accuracy and the relationship between online and offline events during the study period are compromised. Consequently, the identified correlations could vary if it were possible to limit the data more accurately to the analyzed period.

Another limitation of the study is the partial loss of visual material during data scraping or downloading from the social network. Forty-one percent of the images were disabled at the time of download due to IPs blocks, CAPTCHAs, outdated templates, or other factors. While this does not affect the overall interpretation of digital discourse, it reduces the possibility of examining systematic patterns in memes, icons, or photographs, which could provide relevant information about emotional content or rhetoric.

It is important to note, however, that the study focused on actions and practice groups located in metropolitan Lima because the phenomenon unfolded primarily in the capital city. This geographical focus limits the ability to generalize the findings to the national level since it cannot be assumed that the dynamics observed in Lima fully reflect the rest of the country's reality. Therefore, conservative mobilizations in other regions may exhibit different characteristics depending on the cultural, socioeconomic, and political nuances of each local context and how they are connected to the central process that occurred in the capital.

Finally, despite being eight years old at the time of the study, the validity of the database used is worth highlighting. While this time lag may create limitations for more detailed longitudinal analyses, the collected digital material remains significant in terms of its richness and relevance. The data provides a deeper understanding of the origins, dynamics, and forms of articulation of a conservative movement that currently has a significant impact on the Peruvian political landscape and, more broadly, the Latin American context.

6. CONCLUSIONS

The emergence of the conservative movement in Peru, particularly during its phase of mass mobilization against the gender-focused school curriculum, can be understood as the formation of a social movement that arose in response to the advancement of policies aimed at expanding citizens' rights. This movement deploys connective action and progressively evolves into a political actor capable of influencing public debate and state institutions. Simultaneously, it represents a sociocultural phenomenon characterized by a conservative moral narrative that reacts to structural transformations in family life, identity politics, and normative citizenship frameworks in contemporary Peru.

Conservative mobilization manifests itself as a movement articulated through connective action and simplified, emotionally charged ideological discourses centered on a normative vision of the "natural" family—heterosexual, nuclear, and patriarchal—as a fundamental pillar of society. Based on this premise, there is an explicit rejection of the expansion of individual civil rights regarding gender, sexuality, and diversity. Simplified rhetoric and polarizing discourse allow for high communicative effectiveness, facilitating dissemination among diverse social sectors, both online and offline, through a homogeneous, ideologically positioned, exclusionary narrative. The radicalization of messages, the personalization of publications, and the strategic use of identity symbols and media figures have contributed to the consolidation of a cohesive yet deeply polarized and polarizing digital community. This radicalization expresses not only a cultural or moral reaction to educational policies perceived as threatening but also reveals broader ideological transformations in significant sectors of the population where moral conservatism is reconfiguring itself as an active political identity.

As a form of digital political activism, connective action enables the spread of simplified, emotionally charged ideological discourses, such as the anti-gender narrative described above, among individuals who may not be formally organized but connect through the exchange of symbolically resonant content. This new collective identity of social groups that share a denial of diversity and a sustained affective polarization consists of phrases and slogans, confrontational videos, and the exaltation of media recognition as an achievement.

Based on these findings, several relevant lines of research emerge. First, new comparative analyses must be conducted to evaluate how this and other contemporary hybrid mobilizations—whether conservative or progressive—shape public discourse and foster political influence through digital platforms and connective actions. Second, future research could explore the role of algorithms in amplifying polarizing content and the long-term consequences of affective polarization on national social cohesion and democratic governance. Finally, it is crucial to debate the responses that the state and organized civil society can offer to these concerted forms of connective action, which jeopardize the implementation of laws and social policies. The best first step to immunize citizens against the restrictions on pluralism and informed debate that mobilizations like the one studied seek to impose would be the widespread integration of critical digital literacy components into formal educational programs.

7. REFERENCES 

Abas, J. M., Baui, L. Z., & Bondoc, A. (2025, May 22). Analyzing social media sentiments and progression of emotions over time on strategic gaps in video-on-demand digital streaming platforms. SSRN. https://doi.org/10.2139/ssrn.5277948

Acosta, M. (2019). Campañas electorales y plataformas digitales. La precandidatura de Alberto Fernández en Facebook. Question/Cuestión, 1(64). https://doi.org/10.24215/16696581e208

Ayers, J. W., Chu, B., Zhu, Z., Leas, E. C., Smith, D. M., Dredze, M., & Broniatowski, D. A. (2021). Spread of Misinformation About Face Masks and COVID-19 by Automated Software on Facebook. JAMA Internal Medicine, 181(9), 1251-1253. https://doi.org/10.1001/jamainternmed.2021.2498

Bakardjieva, M. (2023). The other civil society: Digital media and grassroots illiberalism in Bulgaria. European Societies, 25(2), 304-325. https://doi.org/10.1080/14616696.2021.1966072

Bárcenas Barajas, K. (2022). Movimientos antigénero en América Latina: Cartografías del neoconservadurismo. Universidad Nacional Autónoma de México.

Barnes, N. (2022). LGBTQ Human Rights and Conservative Backlash: A Case Study of Digital Activism in Mexico. International Journal of Arts, Humanities and Social Sciences, 3 (2), 1-11. https://ijahss.net/assets/files/1644688503.pdf

Benkler, Y., Farris, R., & Roberts, H. (2018). Network Propaganda: Manipulation, Disinformation, and Radicalization in American Politics. Oxford University Press. https://doi.org/10.1093/oso/9780190923624.001.0001

Bennett, W. L., & Livingston, S. (2025). Platforms, politics, and the crisis of democracy: Connective action and the rise of illiberalism. Perspectives on Politics, 1-20. https://doi.org/10.1017/S1537592724002123

Bennett, W. L., & Segerberg, A. (2012). The Logic of Connective Action. Information, Communication & Society, 15(5), 739-768. https://doi.org/10.1080/1369118X.2012.670661

Blee, K. M. (2017). How the study of white supremacism is helped and hindered by social movement research. Mobilization: An International Quarterly, 22(1), 1-15. https://doi.org/10.17813/1086-671X-22-1-1

Bob, C. (2012). The Global Right Wing and the Clash of World Politics. Cambridge University Press. https://doi.org/10.1017/CBO9781139031042

Brito, A. S. (2020). Disputando la opinión pública digital: #ConMisHijosNoTeMetas y los colectivos LGBTI en Ecuador. GIGAPP Estudios Working Papers, 7(166-182), Article 166-182. https://www.gigapp.org/ewp/index.php/GIGAPP-EWP/article/view/209

Carrera Walling, S. K. (2021). ¿Persuasión o manipulación? Con mis hijos no te metas: análisis crítico de su discurso en Facebook [Tesis]. UNMSM. https://cybertesis.unmsm.edu.pe/handle/20.500.12672/17300

Castro-Pérez, R. (2022). Performing Conservatism. A Study of Emerging Political Mobilisations in Latin America using “Social Media Drama” Analysis. In E. Costa, P. G. Lange, N. Haynes J. Sinanan (Eds.), The Routledge Companion to Media Anthropology (1.ª ed., pp. 465-478). Routledge. https://doi.org/10.4324/9781003175605-45

Castro-Pérez, R. (2024). La “batalla mediática” de Padres En Acción y el discurso conservador en la escena política peruana. Letras (Lima), 95(141), 227-244. https://doi.org/10.30920/letras.95.141.14

Chagas, V. (2022). WhatsApp and Digital Astroturfing: A Social Network Analysis of Brazilian Political Discussion Groups of Bolsonaro’s Supporters. International Journal of Communication, 16, 2431-2455. https://ijoc.org/index.php/ijoc/article/view/17296

Chomel, V., Panahi, M., & Chavalarias, D. (2023). Manipulation During the French Presidential Campaign: Coordinated Inauthentic Behaviors and Astroturfing Analysis on Text and Images. In H. Cherifi, R. N. Mantegna, L. M. Rocha, C. Cherifi, y S. Miccichè (Eds.), Complex Networks and Their Applications XI (pp. 121-134). Springer International Publishing. https://doi.org/10.1007/978-3-031-21127-0_11

Collins, J. (2024). The gift of Gab: A netnographic examination of the community building mechanisms in far-right online space. Terrorism and Political Violence, 37(3), 349-368. https://doi.org/10.1080/09546553.2023.2296514

Cuesta-Cambra, U., Martínez-Martínez, L. y Niño-González, J.-I. (2019). An analysis of pro-vaccine and anti-vaccine information on social networks and the internet: Visual and emotional patterns. Profesional de La Información, 28(2). https://doi.org/10.3145/epi.2019.mar.17

Cuevas-Calderón, E., Dongo, E. Y., & Kanashiro, L. (2022). Conservadores en TikTok: Polarización social en el Perú. Revista Prisma Social, 39. https://revistaprismasocial.es/ps/article/view/4865

Davidson, T. (2021). Audience Engagement and the Dynamics of Online Activism: Far-Right Mobilization on Facebook [Preprint]. SocArXiv. https://doi.org/10.31235/osf.io/rwfjc

Doroshenko, L., & Tu, F. (2023). Like, Share, Comment, and Repeat: Far-right Messages, Emotions, and Amplification in Social Media. Journal of Information Technology & Politics, 20(3), 286-302. https://doi.org/10.1080/19331681.2022.2097358

Duárez Mendoza, J. L. (2024). El conservadurismo radical en Perú. El Caso de Renovación Popular (2020- 2024). Letras (Lima), 95(141), 73-92. https://doi.org/10.30920/letras.95.141.5

Featherstone, J. D., Barnett, G. A., Ruiz, J. B., Zhuang, Y., & Millam, B. J. (2020). Exploring childhood anti-vaccine and pro-vaccine communities on twitter – a perspective from influential users. Online Social Networks and Media, 20, 100105. https://doi.org/10.1016/j.osnem.2020.100105

Fejős, A., Zentai, V., Bladini, M., Igareda, N., Helmstad, A., Pascale, A. A., Schulz, A., Sette, R., Sicurella, S., Svensson, E.-M., & Wilhelm, C. (2021). Anti-Gender Hate Speech in Populist Right-Wing Social Media Communication. https://ddd.uab.cat/record/250923

Ferré Pavia, C., & Sambucetti, M. F. (2022). El neoconservadurismo religioso en Twitter: La campaña #ConMisHijosNoTeMetas y el discurso contra la igualdad de género. Teknokultura: Revista de Cultura Digital y Movimientos Sociales, 19(1), 55-66. https://doi.org/10.5209/tekn.75368

Figeac, J., Smyrnaios, N., Salord, T., Cabanac, G., Fraisier, O., Ratinaud, P., & Seffusatti, F. (2020). Information-sharing practices on Facebook during the 2017 French presidential campaign: An “unreliable information bubble” within the extreme right. Communications, 45(s1), 648-670. https://doi.org/10.1515/commun-2019-0193

Fraxanet, E., Kaltenbrunner, A., Germano, F., & Gómez, V. (2025). Analyzing news engagement on Facebook: Tracking ideological segregation and news quality in the Facebook URL dataset [Preprint]. arXiv. https://arxiv.org/abs/2409.13461

Freelon, D., Marwick, A., & Kreiss, D. (2020). False equivalencies: Online activism from left to right. Science, 369(6508), 1197-1201. https://doi.org/10.1126/science.abb2428

Gallagher, R. J., Reagan, A. J., Danforth, C. M., & Dodds, P. S. (2018). Divergent discourse between protests and counter-protests: #BlackLivesMatter and #AllLivesMatter. PLOS ONE, 13(4), e0195644. https://doi.org/10.1371/journal.pone.0195644

Gerbaudo, P., De Falco, C. C., Giorgi, G., Keeling, S., Murolo, A., & Nunziata, F. (2023). Angry Posts Mobilize: Emotional Communication and Online Mobilization in the Facebook Pages of Western European Right-Wing Populist Leaders. Social Media + Society, 9(1), 205630512311633. https://doi.org/10.1177/20563051231163327

Ghani, A., & Rahmat, H. (2023). Confirmation Bias in Our Opinions on Social Media: A Qualitative Approach. Journal of Communication, Language and Culture, 3(1), 47-53. https://doi.org/10.33093/jclc.2023.3.1.4

Giglietto, F., Marino, G., Mincigrucci, R., & Stanziano, A. (2023). A Workflow to Detect, Monitor, and Update Lists of Coordinated Social Media Accounts Across Time: The Case of the 2022 Italian Election. Social Media + Society, 9(3). https://doi.org/10.1177/20563051231196866

Giglietto, F., Righetti, N., Rossi, L., & Marino, G. (2020a). Coordinated Link Sharing Behavior as a Signal to Surface Sources of Problematic Information on Facebook. International Conference on Social Media and Society, 85-91. https://doi.org/10.1145/3400806.3400817

Giglietto, F., Righetti, N., Rossi, L., & Marino, G. (2020b). It takes a village to manipulate the media: Coordinated link sharing behavior during 2018 and 2019 Italian elections. Information, Communication & Society, 23(6), 867-891. https://doi.org/10.1080/1369118X.2020.1739732

Graham, T., Bruns, A., Zhu, G., & Campbell, R. (2020). Like a virus: The coordinated spread of coronavirus disinformation [Report]. Centre for Responsible Technology. https://apo.org.au/node/305864

Grover, T., & Mark, G. (2019). Detecting Potential Warning Behaviors of Ideological Radicalization in an Alt-Right Subreddit. Proceedings of the International AAAI Conference on Web and Social Media, 13, 193-204. https://doi.org/10.1609/icwsm.v13i01.3221

Gruzd, A., Mai, P., & Soares, F. B. (2022). How coordinated link sharing behavior and partisans’ narrative framing fan the spread of COVID-19 misinformation and conspiracy theories. Social Network Analysis and Mining, 12(1), 118. https://doi.org/10.1007/s13278-022-00948-y

Ham, C.-D., Lee, J., Hayes, J. L., & Bae, Y. H. (2019). Exploring sharing behaviors across social media platforms. International Journal of Market Research, 61(2), 157-177. https://doi.org/10.1177/1470785318782790

Iyengar, S., Lelkes, Y., Levendusky, M., Malhotra, N., & Westwood, S. (2019).  The Origins and Consequences of Affective Polarization in the United States. Annual Review of Political Science, 22, 129-146. 
https://doi.org/10.1146/annurev-polisci-051117-073034

Jamieson, K., & Capella, J. (2010). Echo Chamber. Rush Limbaugh and the Conservative Media Establishment. Oxford University Press.

Kakavand, A. E. (2023). Far-right social media communication in the light of technology affordances: A systematic literature review. Annals of the International Communication Association, 48(1), 37-56. https://doi.org/10.1080/23808985.2023.2280824

Kasimov, A. (2023). Decentralized hate: Sustained connective action in online far-right community. Social Movement Studies, 0(0), 1-19. https://doi.org/10.1080/14742837.2023.2204427

Keller, F. B., Schoch, D., Stier, S., & Yang, J. (2020). Political Astroturfing on Twitter: How to Coordinate a Disinformation Campaign. Political Communication, 37(2), 256-280. https://doi.org/10.1080/10584609.2019.1661888

Kessler, G., Vommaro, G., & Paladino, M. (2021). Antipopulistas reaccionarios en el espacio público digital. Estudios Sociológicos de El Colegio de México, 40(120). https://doi.org/10.24201/es.2022v40n120.2213

Kim, C., & Yang, S.-U. (2017). Like, comment, and share on Facebook: How each behavior differs from the other. Public Relations Review, 43(2), 441-449. https://doi.org/10.1016/j.pubrev.2017.02.006

Kim, J. W. (2018). They liked and shared: Effects of social media virality metrics on perceptions of message influence and behavioral intentions. Computers in Human Behavior, 84, 153-161. https://doi.org/10.1016/j.chb.2018.01.030

Kim, S., & Kim, J. (2023). The Information Ecosystem of Conspiracy Theory: Examining the QAnon Narrative on Facebook. Actas de la ACM sobre interacción persona-computadora, 7(CSCW1). https://doi.org/10.1145/3579626

Knüpfer, C. B., Yang, Y., & Cowburn, M. (2025, Septiembre 24). The logic of connective faction: How digitally-networked elites and hyper-partisan media radicalize politics. OSF Preprints. https://doi.org/10.31219/osf.io/2exfz_v3

Knüpfer, C., & Klinger, U. (2025). In-groups and outrage: How narratives and affect shape digital surrogate networks and radicalize right-wing parties. In S. Livingston, & M. Miller (Eds.), Connective action and the rise of the far-right: Platforms, politics, and the crisis of democracy (Cap. 5). Oxford University Press. https://doi.org/10.1093/oso/9780197794937.003.0005

Knüpfer, C., Hoffmann, M., & Voskresenskii, V. (2022). Hijacking MeToo: Transnational dynamics and networked frame contestation on the far right in the case of the ‘120 decibels’ campaign. Information, Communication & Society, 25(7), 1010-1028. https://doi.org/10.1080/1369118X.2020.1822904

Koltai, K. S. (2020). Human values and scientific controversies: Studying vaccine information behavior on social networking sites [Tesis]. The University of Texas at Austin. https://doi.org/10.26153/tsw/12582

Legua Aranibar, Y. E. (2021). Ministerio de Educación y con mis hijos no te metas: Disputa comunicativa y opinión pública sobre el currículo escolar 2017. USIL Repositorio Institucional. https://hdl.handle.net/20.500.14005/11620

Lerman, K., Feldman, D., He, Z., & Rao, A. (2024). Affective polarization and dynamics of information spread in online networks. npj Complexity 1, 8. https://doi.org/10.1038/s44260-024-00008-w

López, J. A. (2020). Aborto, contramovilización y estrategias de comunicación contra la expansión de derechos en México. Revista interdisciplinaria de estudios de género de El Colegio de México, 6. https://doi.org/10.24201/reg.v6i0.621 

Lukito, J., Yang, Y., & Kim, S. J. (2025). How QAnon developed from a fringe group to a digital surrogate for the GOP. In S. Livingston, & M. Miller (Eds.), Connective action and the rise of the far-right: Platforms, politics, and the crisis of democracy (Cap. 3). Oxford University Press. https://doi.org/10.1093/oso/9780197794937.003.0003

Luriaga, R. L., Samson, B. D., Mama, B. R. Z. I., Ebardo, R. A. (2025). LDA-Based Topic Modelling on COVID-19 FLiRT Variant in Social Media. In H. A. Le Thi, H. M. Le, & Q.T. Nguyen (Eds.), Advances in Data Science and Optimization of Complex Systems. ICAMCS 2024. Lecture Notes in Networks and Systems (1569). Springer, Cham. https://doi.org/10.1007/978-3-032-00267-9_11

Marwick, A., Clancy, B., & Furl, K. (2022). Far-Right Online Radicalization: A Review of the Literature. The Bulletin of Technology & Public Life. https://doi.org/10.21428/bfcb0bff.e9492a11

McIvor, M. (2019). Human Rights and Broken Cisterns: Counterpublic Christianity and Rights-based Discourse in Contemporary England. Ethnos, 84(2), 323-343. https://doi.org/10.1080/00141844.2017.1420671

McNerney, H. W., Spann, B., Mead, E. L., Kready, J., Marcoux, T., & Agarwal, N. (2022). Assessing the influence and reach of digital activity amongst far-right actors: A comparative evaluation of mainstream and ‘free speech’ social media platforms. For(e)Dialogue, 4(1). https://doi.org/10.21428/e3990ae6.60c47409

Meneses, D. (2019). Con Mis Hijos No Te Metas: Un estudio de discurso y poder en un grupo de Facebook peruano opuesto a la «ideología de género». Anthropologica, 37(42), 129-154. https://doi.org/10.18800/anthropologica.201901.006

Minici,  M., Luceri, L., Cinus, F., & Ferrara, E. (2024). Uncovering Coordinated Cross-Platform Information Operations Threatening the Integrity of the 2024 U.S. Presidential Election Online Discussion. SSRN http://dx.doi.org/10.2139/ssrn.4965049

Muro Ampuero, E. (2022). Neo-Pentecostalism and religious conservatism on Facebook: A digital ethnography in Peruvian Pro-Family and Pro-Life groups. Desde El Sur, 14(2). https://doi.org/10.21142/DES-1401-2022-0017

Páez González, C. L., & Peña García, A. de la. (2018). El discurso sobre género en la página de Facebook del Movimiento Pro Vida y Pro Familia en Paraguay. Chasqui: Revista Latinoamericana de Comunicación, 139, 229-243. https://revistachasqui.org/index.php/chasqui/article/view/3728

Pariser, E. (2017). El Filtro burbuja: cómo la red decide lo que leemos y lo que pensamos. Taurus.

Páucar Villacorta, D. (2025). Data_Accion_Conectiva [Data Set].  Zenodo. https://doi.org/10.5281/zenodo.15368544

Peredo Rodríguez, V. (2022). Organización, discurso y actividad digital. Una aproximación al movimiento antivacunas en grupos de Whatsapp. Punto Cero, 27(44), 71-88. https://doi.org/10.35319/puntocero.202244192

Pérez-Vela, R. (2022). Activismos religiosos neo-progresistas: La narrativa contestataria evangélica en el espacio público peruano. Iztapalapa. Revista de ciencias sociales y humanidades, 43(92), 73-110. https://doi.org/10.28928/ri/922022/atc3/perezvelar

Pond, P., & Lewis, J. (2019). Riots and Twitter: Connective politics, social media and framing discourses in the digital public sphere. Information, Communication & Society, 22(2), 213-231. https://doi.org/10.1080/1369118X.2017.1366539

Recuero, R., Soares, F. B., Vinhas, O., Volcan, T., Hüttner, L. R. G., & Silva, V. (2022). Bolsonaro and the Far Right: How Disinformation About COVID-19 Circulates on Facebook in Brazil. International Journal of Communication, 16, 148-171. https://ualresearchonline.arts.ac.uk/id/eprint/19805/1/17724-61165-2-PB.pdf

Reguer-Petit, M., & Morabito, L. (2017). Les traces de La Manif pour tous. In F. Gougou,, & V. Tiberj(Dir.). La déconnexion électorale. Un état des lieux de la démocratie française. Fondation Jean-Jaurès.

Rim, H., Lee, Y., & Yoo, S. (2020). Polarized public opinion responding to corporate social advocacy: Social network analysis of boycotters and advocators. Public Relations Review, 46(2), 101869. https://doi.org/10.1016/j.pubrev.2019.101869

Rohlinger, D. A., & Bunnage, L. (2017). Did the Tea Party Movement Fuel the Trump-Train? The Role of Social Media in Activist Persistence and Political Change in the 21st Century. Social Media + Society, 3(2), 2056305117706786. https://doi.org/10.1177/2056305117706786

Romero-Rodríguez. L., Castillo-Abdul, B., & Cuesta-Valiño, P. (2023). The Process of the Transfer of Hate Speech to Demonization and Social Polarization. Politics and Governance, 11(2). https://doi.org/10.17645/pag.i349

Ross-Arguedas, A., Robertson, C., Fletcher, R., & Nielsen, R. (2022). Echo chambers, filter bubbles, and polarisation: a literature review. Reuters Institute. https://reutersinstitute.politics.ox.ac.uk/echo-chambers-filter-bubbles-and-polarisation-literature-reviewZ

Schradie, J. (2019). The Revolution That Wasn’t: How Digital Activism Favors Conservatives. Harvard University Press. https://doi.org/10.4159/9780674240438

Sgró-Ruata, C. (2021). Política sexual, activismo conservador y redes sociodigitales. Virtualis, 12(23), 67-94. https://doi.org/10.46530/virtualis.v12i23.384

Shahin, S., & Ng, Y. M. M. (2022). Connective action or collective inertia? Emotion, cognition, and the limits of digitally networked resistance. Social Movement Studies, 21(4), 530-548. https://doi.org/10.1080/14742837.2021.1928485 

Sundar, S. S., Snyder, E. C., Liao, M., Yin, J., Wang, J., & Chi, G. (2024). Sharing without clicking on news in social media. Nature Human Behaviour. https://doi.org/10.1038/s41562-024-02067-4

Tello, K. (2019). Causas de la politización reactiva del conservadurismo evangélico en el Perú contemporáneo. Politai: Revista de Ciencia Política, 19, 13-33. https://doi.org/10.18800/politai.201902.001 

Tenenboim, O. (2022). Comments, Shares, or Likes: What Makes News Posts Engaging in Different Ways. Social Media + Society, 8(4). https://doi.org/10.1177/20563051221130282

Theocharis, Y., Boulianne, S., Koc-Michalska, K., & Bimber, B. (2023). Platform affordances and political participation: How social media reshape political engagement. West European Politics, 46(4), 788-811. https://doi.org/10.1080/01402382.2022.2087410

Thompson, N., Wang, X., & Daya, P. (2020). Determinants of News Sharing Behavior on Social Media. Journal of Computer Information Systems, 60(6), 593-601. https://doi.org/10.1080/08874417.2019.1566803

Törnberg, A., & Törnberg, P. (2023). White supremacists anonymous: How digital media emotionally energize far-right movements. Journal of Information Technology & Politics, 22(1), 131-148. https://doi.org/10.1080/19331681.2023.2262459

Vico, E. A., & Rey, P. R. (2020). Discurso, estrategias e interacciones de Vox en su cuenta oficial de Instagram en las elecciones del 28-A. Derecha radical y redes sociales. Revista Latina de Comunicación Social, 77. https://doi.org/10.4185/RLCS-2020-1455

Xu, W. W. (2020). Mapping Connective Actions in the Global Alt-Right and Antifa Counterpublics. International Journal of Communication, 14(0). https://ijoc.org/index.php/ijoc/article/view/11978

Yoshida, M., Sakaki, T., Kobayashi, T., & Toriumi, F. (2021). Japanese conservative messages propagate to moderate users better than their liberal counterparts on Twitter. Scientific Reports, 11(1). https://doi.org/10.1038/s41598-021-98349-2

Yu, W. E. S. (2022). A Framework for Studying Coordinated Behaviour Applied to the 2019 Philippine Midterm Elections. In X.-S. Yang, S. Sherratt, N. Dey,, & A. Joshi (Eds.), Proceedings of Sixth International Congress on Information and Communication Technology (pp. 721-731). Springer. https://doi.org/10.1007/978-981-16-2380-6_63

Yuan, X., Schuchard, R. J., & Crooks, A. T. (2019). Examining Emergent Communities and Social Bots Within the Polarized Online Vaccination Debate in Twitter. Social Media + Society, 5(3), 2056305119865465. https://doi.org/10.1177/2056305119865465

Żuk, P., & Żuk, P. (2020). Right-wing populism in Poland and anti-vaccine myths on YouTube: Political and cultural threats to public health. Global Public Health, 15(6), 790-804. https://doi.org/10.1080/17441692.2020.1718733

AUTHORS' CONTRIBUTIONS, FUNDING AND ACKNOWLEDGMENTS

Authors' Contributions:

Conceptualization: Castro-Pérez, Raúl. Software: Páucar Villacorta, Diego. Validation: Tejedor, Santiago, and Romero-Rodríguez, Luis. Formal analysis: Castro-Pérez, Raúl and Páucar Villacorta, Diego. Data curation: Castro-Pérez, Raúl and Páucar Villacorta, Diego. Drafting-Preparation of the original draft: Castro-Pérez, Raúl. Drafting, Review, and Editing: Castro-Pérez, Raúl. Visualization: Páucar Villacorta, Diego. Supervision: Tejedor, Santiago, and Romero-Rodríguez, Luis. Project Management: Castro-Pérez, Raúl. All authors have read and accepted the published version of the manuscript: Castro-Pérez, Raúl; Tejedor, Santiago; Romero-Rodríguez, Luis, and Páucar Villacorta, Diego.

Funding: This research received funding from the Seed Fund for Teachers at the Scientific University of the South, with grant number 029-2023-PRO99.

Conflict of interest: The authors of this article declare that there are no conflicts of interest.

 

AUTHORS:

Raúl Castro-Pérez

Scientific University of the South.

Dean of the Faculty of Communication at the Scientific University of the South. Director of the journal Desde El Sur (Scopus Q2) and coordinator of the MediaLab Research Group, a transmedia literacy initiative. He is a doctoral candidate in Communication and Journalism at the Autonomous University of Barcelona and holds a Master's degree in Communication and Society from Goldsmiths, University of London. He received the Alban Program Scholarship for Advanced Studies from the European Union and a Konrad Adenauer Foundation Scholarship. His research has been published in The Routledge Companion of Media Anthropology, edited by Costa, E., Lange, P., et al. (2022), and Latin American Visual Anthropologies, edited by Cánepa, G., Borea, G., et al. (2024). 

ccastroperez@cientifica.edu.pe 

Índice H: 5

Orcid ID: https://orcid.org/0000-0002-8986-0888 

Scopus ID: https://www.scopus.com/authid/detail.uri?authorId=58009243900 

 

Santiago Tejedor

Autonomous University of Barcelona.

He is a professor at the Faculty of Communication Sciences at the UAB and the coordinator of the Doctoral Program in Communication and Journalism. He is also the Director of the Communication and Education Office and of the Master's Program in Communication and Education. He is also the Editor-in-Chief of the journal Anàlisi (Scopus Q1). Tejedor holds Doctorates in Journalism and Communication Sciences from the UAB and in Project Engineering from the Polytechnic University of Catalonia. His research topics include artificial intelligence, online politics, digital journalism, travel, the environment, and media education. His most recent books include La inteligencia artificial en el periodismo (Artificial Intelligence in Journalism) (2024) and Viajar: una curiosa vuelta al mundo del viaje (Traveling: A Curious Journey Around the World of Travel) (2023).

Índice H: 36

Orcid ID: https://orcid.org/0000-0002-5539-9800 

Scopus ID: https://www.scopus.com/authid/detail.uri?authorId=26433952000

Luis M. Romero-Rodríguez

University of Sharjah.

ESAI Business School, Espíritu Santo University

He is a full professor at the University of Sharjah's College of Communication (UAE) and a visiting fellow at ESAI Business School at the Universidad Espíritu Santo (Ecuador). He held a postdoctoral fellowship at the Monterrey Institute of Technology in Mexico from 2018 to 2019. He received his doctorate in communication from the universities of Huelva, Seville, Málaga, and Cádiz (2014) and was awarded the Extraordinary Doctoral Award from the University of Huelva (2015). He also received a master's degree in social communication from the University of Almería (Spain, 2010–2011). Winner of the 2021 Napolitan Victory Awards (The Washington Academy) and the 2020 Roblón Award for Emerging Communication Researcher. His recent books include Manual de gestión de la comunicación corporativa (Corporate Communication Management Manual) and Introducción a las relaciones públicas (Introduction to Public Relations)..

lromero@sharjah.ac.ae

Indice H: 39

Orcid ID: https://orcid.org/0000-0003-3924-1517 

Scopus ID: https://www.scopus.com/authid/detail.uri?authorId=57079106700 

 

Diego Páucar Villacorta

National University of San Marcos. Peru

Sociology graduate from the National University of San Marcos. He is a corporate consultant in computational sociology, data science, and artificial intelligence. CEO of Escape.

Índice H: 1

Orcid ID: https://orcid.org/0000-0003-0176-6406

Scopus ID: https://www.scopus.com/authid/detail.uri?authorId=57201365315 

 

Annex 1.

Data Management and Computational Model for Data Analysis

1.       Data Collection and Processing

After downloading the initial sample, the database underwent a verification and cleaning process. Six hundred duplicate publications resulting from the integration of multiple sources were removed, leaving 6,021 valid publications. The publications were then temporally classified into five key periods: (1) before the marches, (2) around the first march, (3) between marches, (4) around the second march, and (5) after the marches.

The main images associated with each publication were also downloaded. In cases where this was not possible, manual screenshots were taken. Approximately 41% of the publications had display or download errors, significantly reducing the total number of processable images. The valid images were thematically organized using the BERTopic multimodal model. A manual reclassification was then performed, resulting in 16 final categories of visual content. With the help of the Gemini 1.5 Pro model, it was possible to analyze the presence of people and embedded text in the images, allowing for the identification of elements such as identity symbols, visible leadership, and explicit messages incorporated into the graphic elements.

1.   Analysis of "Outcome Variables"

The study focuses on user behavior in response to posts from the selected pages, particularly the level of interaction measured by comments, shares, and overall engagement. The evolution of the main messages was analyzed to observe how their radicalization relates to the increase in interaction volume. Two outcome variables were used for this analysis:

• Type of content shared: posts containing external URLs (indicator of connection with networks outside of Facebook) were distinguished from those composed exclusively of text, images or videos uploaded directly by users.
• Number of shares: The Shares variable was used, indicating the raw number of times content was shared. To moderate the effect of unequal page size, logarithmic normalization was applied based on the number of followers of each page in 2024, according to the following formula:

Cross-sharing between Facebook pages was not considered, as less than 5% of posts were republished by other pages within the same platform, preventing the establishment of significant patterns of direct collaboration.

2.   Analysis of Independent Variables

According to the specialized literature, two groups of independent variables were identified: (1) the type of content shared, which is understood as a reflection of the page's strategy toward its audience; and (2) the number and type of linked external sources as an indicator of the level of radicalization and ideological alignment.

These factors were analyzed by classifying texts and images separately using a machine learning model implemented in a Jupyter Lab environment (Python 3.11). The notebook is publicly available at this link: https://gist.github.com/diegopaucarv/701f3a8b8f91d6a094633c81b99a55a4.

The analyzed texts included those written by the pages, as well as texts extracted from shared URLs when available. An unsupervised topic analysis was applied using BERTopic to generate embeddings with the pre-trained ntfloat/multilingual-e5-base model. This process resulted in 18 thematic categories. Simultaneously, a pre-trained model from the pysentimiento library was used to perform sentiment analysis.

Additionally, nouns and noun phrases were extracted from the texts, grouped, and used to estimate co-occurrences and cosine distances between key entities. These semantic relationships were visualized using embedding clouds, which allowed for the detection of relevant discourse clusterings.

BERTopic first applied unsupervised topicalization with restrictive parameters in each iteration to ensure the internal consistency of the clusters. Then, semi-supervised reclassification was applied in cases where the data did not show clear differentiation.

2.       Data Analysis Model

The model explores patterns of digital content dissemination using a combined approach of descriptive statistics and machine learning techniques. First, associations between variables were examined using the following coefficients: Spearman's rank correlation coefficient (ρ) for continuous variables and Cramer's rank correlation coefficient (ν) for categorical variables. The study was then structured into two main components: a binary classification model to estimate the probability that content has been shared and a regression model to predict the magnitude of support received by disseminated content.

Both models were implemented using the Random Forest classification model, which was optimized through cross-validation with a random search (RandomizedSearchCV) and time divisions (GroupKFold). The analyzed periods were considered grouping units. Categorical variables were converted into binary indicators with special treatment for infrequent categories. Continuous variables were scaled using a statistical algorithm (RobustScaler), which reduces the influence of outliers using the variables' interquartile ranges. Finally, the relevance of the variables in the models was assessed through variations in Gini coefficients, which are based on reduction of impurities and reflect changes in model entropy (a decision tree-specific methodology).

The results indicate that both models performed well. The Random Forest or Decision Tree model showed high precision, recall, and F1-score metrics consistently across all studied periods. Meanwhile, the regression model showed adequate mean squared error (MSE), root mean squared error (RMSE), and mean absolute error (MAE) values, though it had a weaker fit in terms of R², especially during periods of social mobilization. Table 1 in this annex presents the main fit indicators for both models.

Table 1. Evaluation Metrics for RandomForest Models.

Model		Before the marches	1st March	Between Marches	2nd March	After the marches
Classification
	Precision	0.93	0.94	0.97	0.92	0.97
	Recall	0.93	0.94	0.97	0.91	0.96
	f1-score	0.93	0.94	0.97	0.91	0.96
	Support	336.3	123.001	273,486	97.945	397,078
Regression
	Mse	0.000154	0.000128	0.000108	0.000113	0.000156
	Rmse	0.012412	0.011319	0.010396	0.010644	0.0125
	Mae	0.003748	0.004206	0.004275	0.003326	0.004179
	r2	-0.009236	-0.476918	-0.09642	-0.024101	-0.041537

Source: Elaborated by the authors.

1. External Validation

For external validation, the official Facebook URL database provided by Meta and Social Science One was used. A total of 1,038 records containing the words “género”, “familia”, “gay”, “lgtb”, “currículo” e “ideología”("gender," "family," "gay," "LGBT," "curriculum," and "ideology") were extracted for Peru within the timeframe of this analysis. The texts and websites were grouped using the BERTOPIC application, a topic modeling tool, and manually, respectively. These results confirmed the consistency of the main findings and strengthened the robustness of the results presented in this study.

2. Details of the Machine Learning Model

For the Decision Tree model, a Random Forest Classifier from the scikit-learn (sklearn) library was used for the binary classification task. The second process used the regression model from the same family. The variables' importance in both models was estimated by the difference in Gini importance statistics (impurity decay), which measures entropy differences within the model. The assembly algorithm splits the sample for each variable and estimates a random number of features to find the best sample split using decision rules (bagging technique). The final result is a list of variables with different "importances." Specifically, the classification model estimates importances  using:

 

And the regression model:

In the first equation, represents a category of categories in a categorical variable, and is the proportion of the sample that belongs to this category. In the second equation, is the total sample size, is the value of a numerical variable for the data point , and is the predicted value for that same variable (usually an average). In both cases, the impurity decrease is calculated as follows:

Here, is a measure of “impurity” (entropy or Gini impurity), e are left and right nodes, and and is the sample size corresponding to both nodes. Finally, the importances of each variable are summed as follows:

Where the total number of nodes created by the classifier or regressor based on the variable , represents the division between the count (with weights) of the data that arrive at the node between the total sample, and is the decrease of impurities.

The parameters of the RandomForest models (both classification and regression) were optimized using RandomizedSearchCV and used weights to avoid bias created by the size of each page. In addition, the training and prediction of both models was done in five folds or temporal divisions according to the periods under study.

Related articles:

Alonso-Ruido, P., Rodríguez-Gil, A. M., Estévez, I., & Regueiro, B. (2025). Igualdad de género y sexismo: una mirada desde la perspectiva del estudiantado de Ciencias de la Educación. Educar, 61(1), 53-68. https://doi.org/10.5565/rev/educar.2241

Arias Hancco, J. F., & Espinoza Villalobos, L. E. (2024). Empoderamiento Femenino: Una mirada desde la igualdad de género y la empleabilidad en universidades peruanas. European Public & Social Innovation Review, 10, 1-18. https://doi.org/10.31637/epsir-2025-619

Casanoves-Boix, J., García-Lafuente, S., & Pérez-Sánchez, M. (2025). Why is brand equity so important for public healthcare managers? A case study of citizens’ perceptions in Ireland. Vivat Academia, 158, 1-27. https://doi.org/10.15178/va.2025.158.e1639

González Osuna, Y., & Barrera Martínez, E. (2025). Igualdad de género en las ayudas a los proyectos cinematográficos. Revista de Comunicación de la SEECI, 58, 1-24. https://doi.org/10.15198/seeci.2025.58.e925

Solís-Salazar, R. M., Hernández-Contreras, R. G., & Báez-Martínez, M. E. (2024). Incorporación de la igualdad de género en un programa de estudios universitarios en México. Revista Científica Retos de la Ciencia, 8(17), 29-44. https://doi.org/10.53877/rc.8.17.20240101.3 

[1] The available data corresponds to metrics collected up until 2024. Due to access restrictions imposed by Meta, it is not possible to obtain disaggregated information per year. This limits detailed longitudinal analysis of historical page behavior.