Networked Gatekeeping and Networked Framing on Twitter Protests in Mexico about the Ayotzinapa Case

This research focuses on a Twitter protest about the disappearance of 43 students in Mexico in 2014 known as the Ayotzinapa case which used the hashtag #PaseDeLista1al43. Social network analysis, interviews and thematic analyses of 3,616 tweets were conducted to investigate how networked gatekeeping and networked framing processes occurred within this networked public. Results show that on this digital protest, elite and non-elite Twitter users were in charge of gatekeeping activities, collaborating to maintain the discussion of the topic of the missing students. The analysis of frames revealed that the protest sought to: humanize the case emphasizing the life of the students; assign responsibility; and portray itself as a way to take an active role and coping emotionally with the disappearance. This works contributes to the literature on networked publics, combining network analysis with interviews to look beyond the digital footprint of the protest to listen to the perspectives of protesters, providing an insight into the dynamics of gatekeeping and framing within a network. Given the porosity, openness, and permanence of Twitter, results from such collaborative relationship of elite and non-elite voices to protest about an issue are now harder to ignore in the public sphere.

n September 26th of 2014, a group of Mexican students from the Ayotzinapa Rural Teachers' School was on its way to Mexico City to commemorate the anniversary of the mass killing of students by the Mexican government four decades before. However, 43 of the pupils never made it to their destination. Instead, they were attacked by local police, arrested, and disappeared. In the aftermath of the disappearance, the event generated widespread political protests in Mexico and was the focus of debate on social media. This study examines a daily protest that for years has occurred on Twitter about the Ayotzinapa case. This demonstration, known as Pase De Lista 1 al 43 (Roll Call 1 to 43), entails a roll call with the names of the students every day at 10:00 pm. The hashtag #PaseDeLista1al43 is used to invite Twitter users to join in by retweeting the protest.
Twitter has been used to protest against government repression or negligence in authoritarian contexts such as China (Leung & Lee, 2014) and Egypt (Papacharissi, 2014). Mexico is a democracy that celebrates free elections to pick its political leaders, but social and political institutions still maintain authoritarian practices, remainders of over 70 years of nondemocratic regime in the 20 th century. Thus, the possibility for voicing criticism of the government through digital means presents an advantage for a public that considers that spaces available to express dissent are limited. Individuals not usually represented are able to introduce and frame narratives about certain events, maintaining the issues in the public eye. Yet, the commercialization of social media and the contestation for spaces for participation difficult these empowerment processes, furthering the "need to look more closely at how social media shapes these political engagements as well as the possible ways of building more productive avenues for social change" (Nikunen, 2018). The use of the hasthag #PaseDeLista1al43 to protest the 43 students' disappearance resulted in the emergence of a networked public, in which technological affordances informed the practices to distribute information and social acts regarding the Ayotzinapa case on the Twitter platform (boyd, 2010). This study analyzes how networked gatekeeping and networked framing processes occurred within that public.
This investigation found that in the protest about the Ayotzinapa case, gatekeeping activities were in charge of a group of elite and non-elite Twitter users that collaborated to maintain the hashtag in use, and as a result, O 238 Galarza Molina -Networked Gatekeeping and Networked Framing the discussion of the topic of the missing students. This collaboration of elite and non-elite users consolidated prominent frames in the protest. The analysis of frames prevalent revealed that the protest sought to humanize the case underscoring students' lives; it also sought to assign responsibility, blaming the federal government; and portrayed the online demonstration as a way to take an active role on the case and coping with the disappearance. Although Twitter is praised for its immediacy for delivering information and opinions, these findings indicate that Twitter can house a long-lasting public that seeks to challenge dominant political narratives, allowing non-elite users to be gatekeepers and frame the messages in the stream using a specific hashtag. This works contributes to the literature on networked publics, providing an insight into the dynamics of gatekeeping and framing within a network, from the perspective of participants of Twitter protests.

Networked Gatekeeping and Networked Framing
As the dynamics for information flow have evolved in digital environments, scholars have looked to adapt gatekeeping theory and framing theory to account for these changes. Gatekeeping investigates "the process of culling and crafting countless bits of information into the limited numbers of messages that reach people each day" (Shoemaker & Vos, 2009, p. 1). Through this mechanism, media shape social reality (Shoemaker & Vos, 2009). Nonetheless, with social media, this top-down transmission structure has transformed into a networked-based structure, as the news diffusion process "now occurs through the fluid relational interplay between the various incumbent and emergent players in networked gatekeeping process" (Ernste, 2014, p. 13). The public is "empowered with more choices in news content, direct channels to speak to the press, and the ability to mobilize online crowds" (Xu & Feng, 2014, p. 431).
Consequently, in a networked context, actors with non-elite status offline can become influential online. These online networks are structured according to what is called scale-free distribution, where few actors which "have relatively high centrality but who generally have low overall cohesiveness to the rest of the network" exist side-by-side to "many network actors, usually members of the general public that "tend to have relatively low centrality but high cohesiveness to a small number of acquaintances" (Ernste, 2014, p. 16). To understand the dynamics in which elites and nonelites negotiate the flow of information in these new circumstances, scholars have developed the concept of networked gatekeeping, a process where "actors are crowdsourced to prominence through the use of conversational, social practices that symbiotically connect elite and crowd in the determination of information relevancy" (Meraz & Papacharissi, 2013, p. 158).
However, this perspective does not understand network structures as flat and absent of hierarchies (Ernste, 2014, p.14). Xu and Feng (2014) analysis of Twitter interactions between traditional gatekeepers (Twitter accounts of US journalists) and citizens evince a disparity between types of actors. Their results show that networked gatekeeping is pluralist, involving people of different levels of political involvement and background. Nonetheless, citizens that are not politically active possessed limited political power in the network, having fewer followers and fewer messages retweeted.
Similarly, traditional framing theory focuses only on how news reports' portrayals of an issue influence audiences (Jiang, Leeman & Fu, 2016), without accounting for the role of audiences in framing construction. Framing theory claims that the way that news reports characterize an issue will influence how the audience understands that issue (Scheufele & Tewksbury, 2007). A frame refers to the selection of certain aspects of perceived reality to be made more salient in a text by defining a problem, its moral evaluation, causes, and solutions (Entman, 1993). Political elites use certain frames about issues to neutralize or render unimportant alternative interpretations of issues (Neuman, Guggenheim, Jang & Bae, 2014).
However, changes in the media landscape have impacted power dynamics between newsmakers and news consumers, who can now participate in framing an issue (Xu & Feng, 2014). Scholars developed the concept of networked framing, "that relies on the interactions between elite and nonelite users and algorithmic aggregations afforded by new digital platforms" (Jiang, Leeman & Fu, 2016, p. 97). This concept takes into account the relationship between elite actors and crowds to bring salience to issues (Meraz & Papacharissi, 2013), as well as the impact that algorithmic aggregations can have in bringing a frame to prominence (Jiang et al., 2016), for example, with hashtags or retweets. Moreover, social media users can revise the frames shared by original posters, going beyond the mere replication of an original message, to advance other communicative functions such as the analysis, interpretation, sharing opinions, and expression of emotion about the issues in question (Nip & Fu, 2016). For instance, research shows that the same themes about the refugee crisis in Finland were framed differently on social media in terms of its more pronounced focus on emotional aspects, affective words, and incivility with respect to traditional media (Pöyhtäri, Nelimarkka, Nikunen, Ojala, Pantti & Pääkkonen, 2019).

#PaseDeLista1al43 as a Networked Public
The protest #PaseDeLista1al43 is led by a famous TV producer in Mexico, Epigmenio Ibarra, who every night at 10 pm posts individual tweets with the face and name of each of the 43 missing students, along with hashtags claiming for justice. These 43 tweets get retweeted each night by the followers of this online demonstration. Besides, as a prelude to invite people to join the roll call, he and other followers utilize the hashtag #PaseDeLista1al43 to discuss the case and other situations that are considered grievances caused by the government. Although the protest has evolved over time, this hashtag is still active. In this study, the use of this hashtag by this group of people is considered as giving way to networked publics, publics that are restructured by networked technologies, which at the same time constitute a space and an imagined collective (boyd, 2010). As such, the aim of this work is to explore the processes of networked gatekeeping and networked framing within this public. The following research questions are posed: RQ1. What are the structural characteristics of the #PaseDeLista1al43 Twitter networked public? RQ2. How does the process of networked gatekeeping unfold in the #PaseDelista1al43 Twitter networked public? RQ3. What are the frames used in the #PaseDeLista1al43 Twitter network?

International and Multidisciplinary Journal of Social Sciences, 8(3) 241
This study took a mixed methods approach to respond to the research questions above. Tweets using the hashtag #PaseDeLista1al43 were acquired from the company Follow The Hashtag; additionally, I conducted interviews with participants of the protest. A total of 166,927 tweets that contained the term "PaseDeLista1al43" were collected. To narrow down this amount, I selected five different moments from January 2015 (origin of the hashtag) to September 2017, using a purposive sampling approach, taking some of the most material moments throughout the protest to observe information-rich episodes of this demonstration. Each moment includes a seven-day range of tweets beginning at a specific starting point. Moment One was the beginning of the use of #PaseDeLista1al43 hashtag, on January 15, 2015 (01/15/2015 -01/21/2015). Even though the event of the disappearance occurred a few months before (on September 2014) and the Twitter protest began around November 2014, the use of this specific hasthag started on January 15, 2015. Before that, other similar competing hashtags were used (e.g., #Los43, #nosfaltan43), until #PaseDeLista1al43 was selected as representative of the Twitter protest by the protest organizer Epigmenio Ibarra. A total of 1,071 tweets were posted during that week. Moment Two is marked by a press conference by the Mexican government presenting the results of an investigation of the students' case on January 27, 2015 (01/27/2015-02/02/2015). At this press conference, the Mexican Attorney General reported that the students had been killed and burned in a dumpyard by a drug gang, and argued that these findings were the "historical truth." This declaration was controversial and the phrase "historical truth" was commonly criticized in tweets belonging to the roll call. During that week, 533 tweets were posted using the hashtag. Moment Three was the first anniversary of the students' disappearance on September 26, 2015 (09/26/2015 -10/02/2015). In total, 679 tweets were posted using the hashtag during that Moment. Moment Four is the date of the release of the last report of the group of independent investigators of the case developed by the Organization of American States, the Interdisciplinary Group of Independent Experts (GIEI), on April 24, 2016 (04/24/2016 -04/30/2016). A total of 492 tweets used the hashtag during that period. Moment Five is the second anniversary of the disappearance, on September 242 Galarza Molina -Networked Gatekeeping and Networked Framing 26, 2016(09/26/2016-10/02/2016. During that week, 841 tweets were posted. In total, 3,616 tweets were analyzed from the five moments selected. Besides the tweets collected, the analyses for this study were based on interviews with participants of the protest. Fifteen in-depth interviews with protest participants were conducted between 2016 and 2018 via Skype. In total, nine participants identified as male and six identified as female; thirteen interviewees were Mexican, one was Scottish, and one Dominican. Their average age was 50 years old. They were reached via snowball sampling technique. This research followed the Institutional Research Board protocol for the protection of human subjects of the institution of adscription of the author at the time of its execution (University of Missouri). Interviewees were promised anonymity for their participation, which was maintained by omitting the use of their names and being identified as "Interviewee #". Audios and transcripts of these conversations are saved in a password-protected computer accessed only by the researcher.
For RQ1 social network analysis (SNA) was conducted. SNA focuses on relationships, and in the patterns of connections among people, groups, and things that emerge from such relationships (Hansen, Shneiderman, & Smith, 2011). A key axiom in SNA is the idea that structure matters, as it can affect the performance of the actors in the network (Borgatti, Mehra, Brass, & Labianca, 2009). To answer RQ1, the tweets metadata was utilized to map the structure of the networked public with Node XL, an open-source template for Excel to develop network graphs. The current study followed the approach used by prior research on Twitter networked publics (Meraz & Papacharissi, 2012;Papacharissi, 2014). Therefore, the analysis of the tweets using the hashtag #PaseDeLista1al43 focused on the linkage data of two Twitter markers, mentions and retweets, to illustrate the overall links between nodes and to observe which are the most important nodes within the network in terms of how many mentions and retweets they got (Hanneman & Riddle, 2005).
For RQ2 the main goal was to understand the process of networked gatekeeping in the protest. Two steps taken towards that goal were 1) identifying what type of Twitter users participated in the protest and 2) establishing the most popular users in the network according to the amount of retweets and favorites in the tweets they created. First, following procedures from prior research (Jiang et al., 2016;Xu et al., 2014), I conducted a descriptive content analysis of the profiles of all Twitter users (N=908) that utilized the hashtag #PaseDeLista1al43 during the five moments analyzed. Taking into account users self-disclosed affiliations and information in the Bio section of their profile, each user was coded to establish their identity. These identities were coded as: non-elite actors (i.e., regular Twitter users), elite actors (i.e., journalists, media outlets, or celebrities), activists/non-governmental organizations, bloggers/blogs, and bots. Also to address RQ2, a descriptive analysis of frequencies of retweets and favorites was conducted to determine who were the most popular Twitter users in the protest. For each moment analyzed, two lists of 10 Twitter users with the highest number of retweets and favorites in a tweet, respectively, were identified.
I conducted deductive (for RQ2) and inductive (for RQ3) thematic analyses to systematically investigate the themes in the transcripts from interviews and in the tweets. Thematic analysis is a process for encoding qualitative information where the goal is to identify themes from the data (Boyatzis, 1998). Themes were examined using procedures from the constant comparative method (CCM), a technique that allows for the reduction of data through coding and recoding until reaching saturation (Glaser & Strauss, 1967). Memos were used to organize the information and group the themes emerging from these analyses.

Social Network Analysis
This research question addressed the structural characteristics of the Twitter protest as a networked public, focusing on two types of interactions between Twitter users who posted messages with the hashtag #PaseDeLista1al43: mentions and retweets. The analysis was conducted using Node XL. Figure  1 and 2 present the network of mentions over the five moments selected for study. These networks illustrate the instances when a user tagged someone else's Twitter handle to directly address a tweet to that person. The network has 668 vertices (Twitter users), 463 unique edges (connections between Twitter users), and 2,988 edges with duplicates, for a total of 3,451 edges.

Galarza Molina -Networked Gatekeeping and Networked Framing
The density of the network (the percentage of connections that occur in the network out of the potential connections that could be made) is 2%. This low density, in addition to the large amount of edges that are duplicates, indicates that there are primarily a few Twitter users who are communicating with most of the Twitter users involved in the network. Similarly, reciprocity between vertices is 4%, indicating that communication within the network is not back and forth in the form of a conversation between users, but is instead unidirectional. Thus, the overall network is characterized as a broadcast network, where mentions are connected to few particular hubs with little interaction with each other (Jackson & Foucault Welles, 2015).  Nodes are sized according to their out-degree centrality and are colored in a scale from red to blue such that those nodes with lower out-degree are red and those with higher out-degree are blue. Figure 1 and 2 portray two types of influencers or prominent users in the network. Figure 1 emphasizes vertices according to their in-degree centrality, which is the amount of connections pointing inward to a vertex. The center shows a clear main actor in the network, the protest organizer @epigmenioibarra. The other vertex that received attention in the form of mentions is Mexican President Enrique Peña Nieto (@EPN). As for the rest of the vertices, there is a diversity of singers and actors, journalists and media outlets, blogs, and citizens in the network. In turn, Figure 2 shows the same network but emphasizes vertices with higher out-degree levels, which are nodes that have more connections pointing outward (users tagging others, not being tagged). In this case, the main Twitter user is once again, @epigmenioibarra, who composes many tweets inviting others to join the protest. However, the rest of the vertices are mostly non-elite Twitter users 246 Galarza Molina -Networked Gatekeeping and Networked Framing (e.g., @jesus_garcia158, @licledezma, @AlidaFreites_tw). Citizens do most of the mentioning, attempting to gain more attention for the #PaseDeLista1al43 hashtag. Figures 3 and 4 show the network of retweets during the five moments studied. This network has 35 vertices, 135 edges in total, including 12 unique edges and 122 edges with duplicates. The density of this network is 2%, and reciprocity is 0%. Thus, similar to the mentions network, retweet interactions on this network are unidirectional and limited to only some users replicating what a few others posted rather than a network of interactions across all users. As such, the structure of retweets as addressivity markers also has the features of a broadcast network, in which retweets are emerging from few central sources. Figures 3 and 4 emphasize the degree centrality of the main users in the network. These graphs display that the protest organizer @epigmenioibarra received more retweets within the network, but he does not retweet others. As shown in Figure 4, which emphasizes the vertices with higher levels of out-degree centrality, the actor that does more retweeting is @AyotzinapaFeed, a Twitter account dedicated to spreading news about the Ayotzinapa case. Most of the larger vertices in both graphs belong to regular Twitter users.  Nodes are sized according to their out-degree centrality and are colored in a scale from red to blue such that those nodes with lower out-degree are red and those with higher out-degree are blue.

Networked Gatekeeping Processes
Research question 2 sought to understand how different type of users (elite and non-elite crowds) interacted with each other in the gatekeeping process, who were the main gatekeepers in this Twitter protest, and how these interactions led to the development of a crowdsourced elite in charge of maintaining the online #PaseDeLista1al43 conversation.
Type of users. The first step to answer RQ2 was to determine the type of users that participated in the protests. Thus, the 908 Twitter accounts of each user who tweeted at least once during the five moments were categorized. Results show that the majority of the participants were non-elite users, with 81% (n = 734) regular citizens posting with the hashtag. In turn, only a small percentage of the overall participants, 7.7% (n = 70) were elite users. Given the topic in question -a case related to human rights violations-activists or NGOs were expected to have more participation in the protest, but their 248 Galarza Molina -Networked Gatekeeping and Networked Framing presence was overall low throughout the periods studied (n = 60). Blogs dedicated to disseminating information and creating awareness about social issues in Mexico also had minor participation in the digital protest (n = 25). Finally, a small fraction of accounts were bots (n = 19).
Most popular users. The amount of retweets and favorites further explicates the gatekeeping process that took place on this network because these two addressivity markers point to the popularity of certain actors. Table 1 shows the users with the highest number of retweets and favorites in a single tweet. The table presents two lists per moment and two lists for the aggregate of the 5 moments, each one indicating the 10 users with the highest number of retweets and favorites in a tweet. The table only shows the highest number of retweets and favorites an actor received in a tweet, regardless of whether that actor had other tweets with a high number of retweets and favorites. As expected, the producer who leads the protest, @epigmenioibarra, was at the top of most of these lists, and in fact, he had dozens of tweets with hundreds of retweets and favorites throughout the five moments. @epigmenioibarra is an elite user who constitutes the main gatekeeper of this protest. Another user who gained prominence in the #PaseDeLista1al43 protest is @DeniseDresserG, a famous Mexican intellectual with many followers (almost 4 million). @DeniseDresserG received the most favorites and retweets in a single tweet for the five moments analyzed. Other elite users who had popular tweets were radio show host @TapiaFernanda, movie actress @marisolgase, journalist @julioastillero, and intellectual @JohnMAckerman. Popular Mexican news blogs such as @MT_enMEXICO, @AnonymousMEX_ and @GastosPendejos also had prominent tweets. The popularity of these Twitter users illustrates the role of elite actors, particularly of @epigmenioibarra and @DeniseDresserG, in disseminating certain messages in the #PaseDeLista1al43 protest. However, non-elite crowds including non-famous citizens with Twitter accounts also have a role as gatekeepers of this networked public. As shown on Table 1, tweets by nonelite users also had popularity in the form of favorites and were also highly disseminated through retweets. What stands out about non-elite users is that they gradually emerged as recurrent popular users in the discussion. Certain non-elite Twitter users gained and maintained notoriety as the use of the International and Multidisciplinary Journal of Social Sciences, 8(3) 249 hashtag #PaseDeLista1al43 persisted throughout two years. Table 1 shows how non-elite Twitter users reappeared among the most popular users in the network in Moment 3, Moment 4, and Moment 5 (e.g., @hekglez, @alexabreo22). Table 1.

Actors with Highest Number of Retweets and Favorites in a Tweet
Note. * Elite User; ** Blog or Blogger; *** Activist or NGO; No * Citizens.

Actors with Highest Number of Retweets and Favorites in a Tweet (continued)
Note. * Elite User; ** Blog or Blogger; *** Activist or NGO; No * Citizens.

Interactions Between Elites and Crowds
A deductive thematic analysis of interview responses was executed, utilizing concepts from networked gatekeeping theory to identify how the process panned out in this particular protest (See Table 2).
The role of elites. The protest was not the result of a planned strategy by a large group or an activism organization. Instead, it emerged from individual citizens who resorted to Twitter to express their grievances about the sudden disappearance of 43 students. An interaction between an elite Twitter user, Epigmenio Ibarra, and a citizen who was a Twitter novice evolved into a systematic online demonstration. The participant who started the roll call explained how the idea emerged: "In the marches (on the streets) we say the numbers 1 to 43 and we yell 'justice' at the end, and at some point he (Epigmenio Ibarra) said, why don't we invite people to do 43 tweets using the numbers? [...] And at some point he began to set a fixed time, it was around November (2014) when he started doing it systematically at 10 at night" (Interviewee 1).
The informant emphasized that she purposefully sought to get attention from a celebrity in order to disseminate the images of the students' illustrations taken from a group of artists website, Ilustradores con Ayotzinapa. This story of the beginning of the protest illustrates the role that elite users with a large following can have on the success of a conversation about an issue on Twitter. Although some other popular figures have taken part in the online demonstration, in the #PaseDelista1al43 protest, the leadership of Ibarra has been key to the continuation of the use of this hashtag. Beyond the nascence of the roll call, other interviewees often refer to the importance of this elite actor's influence in their involvement on the protest. For protesters on this demonstration, gaining attention from this celebrity solidified their commitment to the cause of Ayotzinapa and specifically to their contribution to #PaseDeLista1al43.
Crowdsourced elites. Some of the most active participants in #PaseDeLista1al43 reported that they gained prominence on Twitter thanks to their contribution to this discussion. In addition to the attention from elites such as Ibarra, the main way in which interviewees assess their prominence is by the amount of Twitter followers they have. Protest participants joined the Twitter conversation organically, with the goal of airing their opinions about the case. However, according to their accounts, they ended up developing expansive connections with people congruent with their views and even became relatively popular actors in the network.
Networked gatekeeping. As committed gatekeepers in this networked public, crowd-sourced elites assumed certain functions within the protest. One important function was making sure the hashtag -and with that the Ayotzinapa issue-maintained an online presence. Keeping this hashtag active is important in light of what they perceive to be systemic attempts to minimize the strength of the hashtag in the social network. Similarly, another way to maintain the hashtag is inviting other users via the mention addressivity marker, so that friends on their network or public figures take part of the conversation.
An additional function that these gatekeepers assume is to curate the topics that are talked about in the conversation, making sure that relevant aspects of the case do not escape public attention. For example, an interviewee mentioned a situation in which she suggested that two students that were assassinated on the night of the disappearance of the 43, should also be remembered in the protest. This participant reached out to the protest organizer to influence the #PaseDeLista1al43 conversation, and the students were added to the roll call. In this way, non-elite and elite Twitter users in the protest interacted to filter and configure the topics discussed in the protest.
Participants not only share opinions about the case but they gather and share new information about it from sources they trust (usually left-leaning media). Some protesters merely attach photos or news links, but others create their own images with editing software and attach them to tweets. Some images include information about coming events, fragments of news stories with screenshots from media websites, others take pieces of information about the case and include their own perspective on the story. Thus, protesters appropriated Internet and social media affordances to have a more active role on the roll call, by acting as curators of news stories about International and Multidisciplinary Journal of Social Sciences,8(3) 253 Ayotzinapa, disseminating news through Twitter, and creating their own content with images from the web. Table 2.

Theme Frame
The role of elites "I was tagging Epigmenio Ibarra who at that time had more than 180,000 followers, right? And Epigmenio Ibarra, without knowing me, as it is the issue with networks, he saw that I uploaded these images and came to send some tweets in which he said 'I invite my followers to retweet this' (Interviewee 1)." "He (Ibarra) sent me a direct message and he said, 'hey, I see that you are not here', back then I was not in Mexico, but then (he said) 'I would like to ask you to continue to help us, if you can do it daily well do it', because he noticed that, in a way, the daily impact was receding. Then I said 'yes' and then I assumed the commitment to say, 'well, I have to do it' (Interviewee 5)." Crowdsourced elites "I had very few followers, I had about 150 followers, now I'm at about 3,000, when I started to be active in this movement, I must be around 3,000 [...] This is as the arms of an octopus, they keep growing, you know a person and that person connects you with another, and that person knows another one, and so it goes (Interviewee 11)." Networked Gatekeeping "There are government people who are blocking the hashtag, to stop them we do new ones, that's where the new ones come from, the new hashtags, so that we do not get bots" (Interviewee 3). "In the roll call of Epigmenio Ibarra, they (2 students killed during the police persecution) were not contemplated [...] But well obviously my impact on Twitter is not compared at all to that which Epigmenio has. So then I told him, well I've been doing the roll call to these two boys who were executed, but I think it would be better if you did it, because it is you who has responses (Interviewee 9)." "I try to find the news of the day about Ayotzinapa and post it before the roll call.

Themes and Frames
In order to elucidate the main frames used to talk about the Ayotzinapa case in the #PaseDeLista1al43 protests, a thematic analysis of tweets was conducted. The three broad themes in the tweets were: the Ayotzinapa case involving the disappearance of 43 students, discussion of the social and political conditions in Mexico, and the Twitter protest using the hashtag #PaseDeLista1al43. These themes were framed to humanize the case and assign responsibility for it; to contextualize the disappearance positioning it as part of a larger problem in Mexico; and to discuss the protest as a way to deal with the case in terms of seeking justice and coping emotionally with the situation (See Table 3).

The Ayotzinapa Case
Humanization of the Case. Tweets on the digital protest emphasize the humanity of the 43 missing students by portraying their physiognomy and names. Messages with drawings of their faces include information about their lives, for instance, features of their personality, their nicknames, celebrations of their birthdays, and details about their family. Students are talked about in endearing terms, as "our brothers" or "our friends". Twitter users express familiarity with the students, arguing that they personally have been affected by their disappearance, with claims such as: "I miss him," "I want to hug him," and "I want to know where he is".
Assigning Responsibility. A consistent frame regarding the case refers to the federal government's responsibility on the disappearance. Tweets reject the official version of what happened to the students -that they were killed by a drug gang-, labeled by Mexican attorney general, Jesús Murillo Karam, as the "historical truth." A counterframe that emerged refers to such official version as the "historical lie." The government is depicted as an adversary, responsible for the disappearance of the students. One of the most repeated phrases in the tweets is "It was the State".

Sociopolitical Conditions in Mexico
Contextualizing the Case. Discussants in the protest often referred to aspects of Mexico's sociopolitical context not directly connected to the Ayotzinapa case. For instance, protesters shared concerns regarding violent conditions in Mexico. More specific tweets addressed the large number of missing people in Mexico besides the students. By mentioning Mexico's sociopolitical context, tweets convey that the disappearance was not an isolated case. Moreover, the #PaseDeLista1al43 hashtag was used to talk about other high profile cases involving government's negligence or repression.

#PaseDeLista1al43 Twitter Protest
Citizens' Role in the Case. Tweets using the hashtag #PaseDeLista1al43 also referred to the Twitter protest itself, underlining the role of online protesters in the Ayotzinapa case: how they are impacted by the disappearance and what they can do about it. These messages addressed different aspects of the daily roll call, for example, when does it occur, invitations for others to join, or tagging celebrities to encourage them to invite people in their network. Other tweets show street demonstrations (e.g., depicting crowds on the streets, protests signs, close-ups of protesters), connecting those with the #PaseDeLista1al43 protest. Thus, for the protesters, tweets of the online protest are embedded in the overall Ayotzinapa cause.
Several tweets focus on the reasons to participate in the #PaseDeLista1al43 conversation. Among the motivations to take part of the protest, the tweets expressed the need to demand justice for the students and show solidarity with the students' parents. Another strong motivation was fighting misinformation about the case and furthering social awareness on the issue. Additionally, participation was prompted by the idea that anyone in the country could be the next victim, and by the need to cope with certain feelings about the case. Tweets convey how the authors suffered personal emotional repercussions due to the disappearance, which motivates them to 256 Galarza Molina -Networked Gatekeeping and Networked Framing continue expressing their feelings. As such, these emotional demonstrations allowed protesters to feel their way into the story: positioning themselves as part of the events and taking action about it. Table 3.

Theme Frame Tweet Exemplar
The Ayotzinapa Case

Humanization of the Case
Assigning Responsibility "@epigmenioibarra to demand justice and remember our brothers, there are not times nor schedules, we will be attentive to #PaseDeLista1al43" (trota mundos, 2015).
"@epigmenioibarra uf a little late I join #PaseDeLista1al43 because I DON'T forget, DON'T let go and I think that it is a historical LIE the investigation (Cadena, 2015)."

Discussion
In networked publics-publics that are restructured by networked technologies (boyd, 2010)-members of the public "can be reactors, (re)makers and (re)distributors, engaging in shared culture and knowledge through discourse and social exchange as well as through acts of media reception" (Ito, 2007, p. 3). This study focused on the characteristics of the networked public emergent in the #PaseDelista1al43 Twitter protest to further understand processes of framing and gatekeeping within that public. First, the analysis focused on relationships formed in the network and the patterns of connections among Twitter protesters. Results from SNA showed that the #PaseDelista1al43 networks formed based on retweets and mentions had low reciprocity (no back and forth conversation) and low density (the Twitter users on the network are connected only to few other users in the same network). These characteristics point to a type of network known as a 258 Galarza Molina -Networked Gatekeeping and Networked Framing broadcast network, in which individuals "connect to a single or a small number of actors for the majority of the flow of information" (Himelboim, Smith, Raine, Shneiderman & Espina, 2017, p. 8). Only a few Twitter users in the network, and one in particular (the protest organizer Epigmenio Ibarra), stand out as hubs of information. This structure is similar to the way that information flows in a traditional mass communication model. Yet, the nature of the network becomes more complex when looking at the other main hubs. Unlike typical broadcast networks, which usually have users such media outlets, pundits, and organizations at the center, in the #PaseDelista1al43 network members of the public acquired prominence in the network. Although power is concentrated in a few hubs, these hubs are not exclusively the usual popular actors (e.g., media outlets, pundits, celebrities) but some of them are regular citizens.
Gatekeeping activities were in charge of a group of elite and non-elite Twitter users that worked in collaboration to maintain the hashtag in use, and thus, the discussion of the topic of the missing students. These dynamics are indicative of networked gatekeeping, in which the diffusion of information and the determination of information relevance occurred in a fluid relational interplay between incumbent and emergent players (Ernste, 2014). The analysis of the type of actors in the network showed that a majority were regular Twitter users, considered as non-elite actors. However, there were elite users involved in the digital protest, a TV producer (the protest leader), intellectuals, actors, radio hosts, journalists, and a politician. Despite the fact that they were outnumbered in terms of the amount of users in the network, the prominence of the elite group was evident considering the number of favorites and retweets they achieved. Yet, non-elite users also achieved prominence as the digital protest persisted through time. As such, on the #PaseDeLista1al43 public, few users with high centrality but who generally have low overall cohesiveness to the rest of the network -elites who rarely retweet others or that have conversations with very few within the network-coexist with many network users, likely members of the general public that may have lower centrality but high cohesiveness to a small number of acquaintances (Ernste, 2014). The gatekeeping process in this networked context is, thus, more plural. People with non-elite status offline can author content that achieves prominence and become influential in the network (Meraz & Papacharissi, 2013).
Nonetheless, the fact that there are opportunities to become prominent in the network for non-elite users did not completely erase differences in hierarchies. These differences were pointed out by participants interviewed about #PaseDeLista1al43 protests. The interviewees underscored the leadership of Epigmenio Ibarra in the creation and maintenance of the protest, and how his celebrity status contributed to preserve the demonstration. Similar to findings from Papacharissi and De Fatima Oliveira (2012) who investigated the protests in Egypt in 2011, this analysis identified that hierarchies still matter in social media protests. For this protest, the daily efforts of its leader, an elite Twitter user, were decisive to maintain the hashtag in use. However, the use of Twitter to protest also allows a 'crowdsourced' form of leadership rewarding committed users who reported and curated information on the network (Papacharissi & De Fatima Oliveira, 2012). From the beginning of #PaseDeLista1al43, these non-elite Twitter users were motivated by the attention they got from an elite user to continue with the daily efforts of retweeting the tweets of the protest. Interviewees claimed to have acquired popularity thanks to their participation on the demonstration, gaining up to thousands of followers. In light of that growth within the #PaseDeLista1al43 network, this crowdsourced elite has attributed itself gatekeeping responsibilities, such as protecting the persistence of the hashtag in response to bots attacks; attracting more elite users to the protest; and curating the information that flows in the network. Notably, this investigation contributes in our understanding of how networked gatekeeping occurs, giving place to a symbiotic relationship between elite and non-elite users. This study looks beyond the digital footprint of the protest and listens to the perspective of protesters, who explained the nature of their participation, their motivations, and the dynamics they perceive within the network.
Moreover, the coexistence of elite and non-elite users in the network resulted in their mutual collaboration for the process of networked framing. Narratives about the Ayotzinapa case in the #PaseDeLista1al43 Twitter protest were articulated by "the networked interactions between elites and non-elites, supplemented by algorithmic aggregations" afforded by Twitter, mainly the use of the hashtag #PaseDeLista1al43 and the use of the retweet feature (Jiang et al., 2016, p. 97). The frames that emerged from these crowd 260 Galarza Molina -Networked Gatekeeping and Networked Framing and elites efforts indicate that the protesters rejected a dominant narrative that the Mexican government sought to promote regarding the 43 students' disappearance (mainly, that the case had been solved already). The analysis of tweets with the hashtag #PaseDeLista1al43 showed that, for people participating on this stream, the government is an adversary whose account of the events they strongly reject. Another important frame that appears on this protest is the humanization of the victims by emphasizing the students' faces and names, features of their personalities, and the suffering of their families. These tweets are imbued with emotions about the missing students and their parents, ranging from affection and love for the victims to sadness and outrage about their situation. Furthermore, the authors of the tweets in the #PaseDeLista1al43 Twitter protest reveal a sense of identification with the students and their disappearance. The protesters argue that, because of the general conditions of violence in Mexico, they could potentially be victims of a crime like the students were.
Thus, the analysis of the frames prevalent in the #PaseDeLista1al43 protest evince an interest of protesters to talk about the case, updates on the investigation, and details about the victims' lives, however, the protesters also talk about themselves and their emotions derived from the disappearance of the 43 students. Instead of just representing an event, the tweets in the #PaseDeLista1al43 protest took content from mainstream media and remix it with opinions and feelings about the Ayotzinapa case. In that way, Twitter users utilizing the #PaseDeLista1al43 hashtag position themselves in the story of the disappearance, as opposed to just remaining witnesses of it. Therefore, the protest constituted an opportunity for these Twitter users to feel their way into the 43 students' story. Notably, the reiteration of certain frames, via retweets and favorites, shows the intensity of the endorsement of those frames (Papacharissi & De Oliveira, 2012). Consequentially, the collaboration of elites and crowds to consolidate frames in this networked environment allows the members of this public to portray the news about the disappearance as a mix of facts, opinions, and emotions (Papacharissi, 2014). Despite academic concern about digital activism becoming shielded to avoid harassment (Nikunen, 2018) and for its limits as a form of merely connective but not collective action (Papacharissi, 2016), this study finds that this group of protesters, which emerged with support of International and Multidisciplinary Journal of Social Sciences,8 (3) 261 Twitter's technological affordances, is capable of collectively constructing frames that were sustained over time on Twitter discourse. This study has some limitations. First, the amount of interviews conducted can be deemed small (N=15). Nonetheless, the overall #PaseDelista1al43 protesters population is potentially small, based on two considerations: 1) the amount of retweets that the students' roll call get daily is about 80-100, 2) the estimation from some participants about how many of them contribute to the protest (from 100 to 200). Additionally, the subjects interviewed represented different types of participants of the protest, for example, protesters that do a roll call of their own besides the one conducted by Ibarra, people that just retweet the original roll call, participants that create their own images to curate the content in the protest, participants that are not Mexicans citizens, and Mexicans in another country. This diversity in the type of participation and the demographic characteristics of the protesters indicate that there is variation in the range of participants interviewed. An additional limitation of these interviews is that they were conducted in different periods of time, which could impact participants' responses. However, responses of the subjects interviewed in the first period coincided with those interviewed later during the project.

Conclusion
Hashtags act as a social awareness stream in which subjectivity and objectivity are combined, representing a user-generated collaborative argument about a certain topic (Papacharissi & De Oliveira, 2012). The hashtag #PaseDeLista1al43, consistently used to discuss the Ayotzinapa case since 2014, has consolidated as a social awareness stream that serves as an online home (Papacharissi, 2016) or a long-running epistemic community (Segerberg & Bennett, 2011) for the protest. The narrative constructed by the #PaseDeLista1al43 networked public represents another layer of the Ayotzinapa event. The stories told about Ayotzinapa on the #PaseDeLista1al43 protest constitute another perspective that stands next to others perspectives about the event, for example: "as it is unfolding on the ground, the event as broadcast via television, the event as depicted through the conventions of newspaper storytelling" (Papacharissi, 2016, p. 316).
Protesters highly value the possibility to portray the Ayotzinapa story in their own terms through Twitter. By participating in the online protest, the protesters embrace the opportunity to challenge the narrative that, in their perception, the government is trying to impose.
Communication scholars must continue to look at digital protests to analyze the implications for the practice of activism and citizenship, given the current media landscape. Importantly for this study, the persistency of the use of the hasthag #PaseDeLista1al43 over several years made the identification of regular participants of the online demonstration possible, in order to get their account about the nature of their participation and their motivations to protest. Their responses reinforced findings from previous literature on the symbiotic relationship between elite and non-elite users and the functions that each have on networked gatekeeping and networked framing processes. Given the porosity, openness, and permanence of Twitter, results from such collaborative relationship of elite and non-elite voices to protest about an issue are now harder to ignore in the public sphere. The potential impact of Twitter networked publics to permeate to mainstream narratives beyond the platform should be further investigated.