Another Violent Protest? New Perspectives to Understand Protest Coverage

This study assesses the relationship between two well‐established sets of frames to better understand the news coverage of massive political protests. By relying on Semetko and Valkenburg’s generic frames and McLeod and Hertog’s protest frames, this study aims to identify whether certain generic frames emphasized in news stories increase the tendency to delegitimize protest movements. To this end, we analyzed the news coverage of Chile’s Estallido Social, a series of massive political demonstrations that developed across the country from October to December 2019. Data for this study come from stories published by Radio Bío Bío, the most trusted news outlet in the country, according to Reuters Institute. By analyz‐ ing a sample of 417 stories, we found the coverage replicated patterns that usually delegitimize protest movements, as many of the stories focused on violent acts and depicted demonstrators as deviant from the status quo. We also found a direct relationship between generic frames and protest frames, in which the presence of the former determines that of the latter. Generic frames provide information about how the news media interpret and package the news, which in turn affects demonstration‐related features that the news media pay attention to. As such, we argue that combining both generic and issue‐specific frames is a helpful approach to understanding the complexities of protest news coverage.


Introduction
The year 2019 will be remembered as a year of mass protests, particularly in Latin America. Chile, Colombia, Bolivia, Honduras, Puerto Rico, Peru, Venezuela, and Ecuador saw their citizens take to the streets to mobilize against governments' decisions, social inequalities, political scandals, and allegations of electoral fraud (Wolff, 2020). These massive protests had vastly different economic and political consequences-for Bolivia, it was the resignation of President Evo Morales; for Chile, it was the process of drafting a new constitution. In each case, the protests captured the attention of national and international news media, who set the tone with which the audiences would perceive these demonstrations.
Protests are means of expression used to show disapproval or objection to something (or someone) by individuals who are powerless and cannot prevent that "something" from happening, from social injustice to political corruption (Turner, 1969). As such, protesting is a tool to draw attention to injustice and wrongdoing affecting a group, one that requires a combination of public sympathy and news media interest to accomplish the group's goals. This complex combination can transform the perception of a protest from a peaceful, persuasive event to a coercive action based on extreme violence.
The news media play a significant role in how the public perceives the protest message, depending on how such message is presented in a story. Research on protest coverage has identified several frames the news media use to make sense of protest movements, from sympathetic portrayals (covering protesters' grievances and demands) to extreme depictions of protests and protesters, presenting those as violent and deviant, respectively (e.g., McCluskey et al., 2009;McLeod, 2007;Shoemaker & Reese, 2014). The "protest paradigm" is probably the most studied frame to explain protest coverage, focusing on the media practice of highlighting destruction and violence by demonstrators, quoting official sources, and marginalizing the core reasons behind the movement (Jiménez-Martínez, 2021).
However, the literature has paid little attention to generic frames in the context of mass protests. Initially proposed by Semetko and Valkenburg (2000), generic frames describe perspectives highlighted in news coverage regardless of the issue at hand. In contrast, issuespecific frames pertain to the news topic under study, analyzing specific characteristics and proposing categories that allow for great specificity and detail (Kozman, 2017). In opposition to generic frames, which can transcend topic limitations, issue-specific frames are subject to topic restraints (de Vreese, 2005).
The present study focuses on one set of issue-specific frames proposed by Hertog and McLeod (2001) to study protest coverage. We observe whether certain generic frames increase the salience of specific protest frames in the coverage of massive protest movements. We do so to respond to Kozman's (2017, p. 780) call to study generic and issue-specific frames together: "Generic frames take more of an interpretive, packaging role that could work in tandem with any issue-specific frame, without taking a stance or defining the problem at hand." By doing so, we want to explore the potential of a mixed-frame approach to better understand the complexity of protest news coverage. We argue that protests should be studied using both sets of frames to observe how the presence of certain frames influences the presence of other frames.
Our study analyzes the news coverage of Chile's Estallido Social, a series of massive political demonstrations that developed across the country from October to December 2019. Like many countries in Latin America in 2019, Chile experienced a social-political crisis that aroused broad citizen support. However, the acts of violence were highly emphasized by the news media and ended up obscuring the social demands. By observing how a highly trusted news outlet in Chile covered the Estallido, we aim to understand the framing strategies used by the news media when covering social transformation in a Latin American country.

Frames and Framing Theory
According to framing theory, the news media select and emphasize the elements they consider more relevant to inform an event (Entman & Rojecki, 1993). These elements directly resonate with the audience's internalized mental schemes, affecting how they understand and evaluate an issue (Goffman, 1974;Shoemaker & Reese, 2014). As proposed by Reese (2001, p. 11), frames are "organizing principles that are socially shared and persistent over time, that work symbolically to meaningfully structure the social world." Frames not only affect how people interpret reality, but also provide journalists and communicators with guidelines and patterns to cover the news (McLeod & Detenber, 1999).
Framing research has developed several typologies for the study of journalism practice, from the thematicepisodic dichotomy to explain big-picture coverage versus immediate, non-contextual reporting (Iyengar, 1991), to generic versus issue-specific frames, focused on explaining how issues are portrayed versus functional descriptions of a specific issue (Entman et al., 2009;Kozman, 2017).
Generic frames refer to the way the news media present or package any issue (Kozman, 2017), allowing for the comparison of different issues (de Vreese et al., 2001). Semetko and Valkenburg (2000) proposed five generic frames that have been profusely studied in the framing literature: conflict, attribution of responsibility, human interest, economic consequences, and morality. While the conflict frame focuses on the conflicts between individuals, institutions, and groups, the responsibility frame attributes responsibility to a certain actor for the cause of an event or problem. The humaninterest frame emphasizes human faces and emotional angles in a story to make it more relatable, and the economic consequences frame highlights the economic effects of an event. Lastly, the morality frame reports an event regarding moral or religious values.
Several studies around the globe have proved the broad applicability of generic frames when studying news coverage (e.g., Camaj, 2010;de Vreese, 2005). However, some scholars have pointed out the need to question the truth of this broad applicability when the diversity of countries' socio-historical contexts is taken into consideration (e.g., Aruguete & Koziner, 2014). For example, Aruguete (2010) proposed two additional frames after analyzing the presence of generic frames in Argentina's news: conflict with human impact and conflict resolution. The latter is supported by Gronemeyer and Porath (2017), for whom the context of the study might suggest the need for local adaptations to the original set of five generic frames. For instance, when studying the presence of generic frames in the elite press, Gronemeyer et al., (2020) identified two new frames: defense or damage control frame, which refers to mitigating the damage caused by someone, and the informative frame, centered in providing objective information and data in a neutral tone.
In contrast to generic frames, issue-specific frames are pertinent only to specific topics or events (de Vreese, 2005) and reveal aspects of those topics selected or left out of the coverage (Kozman, 2017). An example of issuespecific frames is protest frames, developed to understand how the news media cover protest movements (Chan & Lee, 1984;McLeod & Hertog, 1999). Studies have shown the news media tend to ignore social movements, especially in the early stages (McLeod, 2007;Shahin et al., 2016;Shoemaker, 1982), but once their actions become more disruptive, the news media pay attention and center their coverage on the violence, portraying the movement from an antagonistic perspective . This pattern has been studied profusely, showing that protest movements are more likely to be covered negatively, depicting protests as violent and protesters as deviant, which delegitimizes the group and its demands (McCluskey et al., 2009;McLeod & Hertog, 1992).

Research on Protest Frames
McLeod and Hertog's typology of protest frames (Hertog & McLeod, 2001;McLeod & Hertog, 1999) includes five framing categories to cover massive demonstrations: confrontation, riot, spectacle (or circus/carnival), debate, and protest. The confrontation frame focuses on the conflicts between protesters and the police, the government, and even the news media. The riot frame emphasizes the violence of protesters and the damages they cause to society, such as lootings and harm to public property. The spectacle frame puts an accent on the oddity, the drama, and the large-scale impact of the demonstrations. The debate frame is probably the only "positive" portrayal in Hertog and McLeod's typology, as this frame covers different points of view and takes distance from the criminalization of protest movements. Lastly, the protest frame focuses on the event of the protest itself. It makes detailed descriptions from beginning to end, describing protesters' activities, police actions, and basically the dos of everyone involved. However, in subsequent studies, Hertog and McLeod (2001) pointed out the protest frame made very few appearances in the news coverage and, as such, it has not been consistently included in protestframe studies.
But regardless of the number of categories, protest frames can impact how audiences perceive a protest and how they react to it . For example, the presence of confrontation and riot provokes more criticism toward protesters and fewer complaints against the police, even during episodes of police brutality. On the other hand, when the coverage legitimizes the protest-as is the case in the debate frame-there is greater identification with protesters and more support for their demands . Several authors (e.g., Harlow et al., 2020;Kilgo et al., 2018;) have adopted the protest frame typology with some modifications. Instead of analyzing the five frames proposed by Hertog and McLeod, recent studies only use four of the five framing typesconfrontation, riot, spectacle, and debate. As Kilgo et al. (2018) explain, the most common frames presented in the news are the ones that marginalize the protest (confrontation, riot, and spectacle), while the ones that aim to legitimize are not so frequent in the mainstream media (debate and protest). However, they mention that the debate frame gives a more substantial voice and space for protesters to express their grievances and demands. Therefore, the "protest" frame has been elided in this new typology proposal.
Recent studies provide nuance to the delegitimization trend in protest coverage. While spectacle is the most common frame when covering protests, Kilgo (2020) points out that the reason behind a protest determines the presence of other frames. For example, the riot frame shows up in racial issues, while the spectacle frame goes together with the debate frame in the coverage of gender, health, and environment-related protests. This is consistent with findings from studies describing the predominance of the spectacle frame, and the incidence of social causes in how the public perceives demonstrations. Nevertheless, some researchers have observed the prominence of other frames over spectacle, like riot or confrontation . Harlow et al. (2020) add that the protest's geographical context also determines what frames are more prominent. In the case of Latin America, the spectacle frame is most common, but the debate frame also shows up frequently, probably because of changes in how social movements are perceived by the public (Harlow et al., 2020). Additionally, news media ideology, and how close to the government the media are, could also impact the coverage-the closer the media are to the authorities, the more they will have to stick to riot and confrontation portrayals (Shahin et al., 2016).

A Mixed-Framing Approach to Understanding Protest Coverage
This study follows Kozman's (2017) approach to framing research, which advocates for the simultaneous study of generic and issue-specific frames to understand how the news media portray issues of public interest. Kozman (2017)-and also Brüggemann and D'Angelo (2018)explored the possibility of jointly applying different framing approaches when conducting framing research. Scholars have pointed out a lack of consistency when it comes to identifying frames, with studies proposing their own framing typologies for each new topic they observe (Cacciatore et al., 2016;Matthes & Kohring, 2008;Scheufele & Tewksbury, 2007,). To address this issue, Kozman (2017) suggests not creating new typologies but to integrate existing ones, particularly generic and issue-specific frames, to complement their functions and reach the specificity needed.
The present study draws upon Kozman's call and aims to examine the nature of the relationship between generic and protest frame typologies. Generic frames present a general perspective of journalistic values when portraying protest news. In contrast, protest frames (as issue-specific frames) highlight specific characteristics and elements of massive demonstrations in relationship to selection and salience. As studies combining both sets of frames are scarce, the literature does not suggest whether generic frames precede and determine issuespecific frames or vice versa. However, generic frames reflect journalistic norms and routines-when journalists pack the news from a certain perspective, those packages teach us about journalism and the worldview of news producers. Thus, we argue that generic frames precede issue-specific frames, and therefore affect which issue-specific frames will be emphasized in the coverage. For instance, violent demonstrations might be addressed using moral values, economic consequences, or attribution of responsibilities. And each approach will lead to different news angles. Then, this study inquires about which generic frames are more impactful in the emergence of protest frames. As such, we ask: RQ: Which generic frames increase the salience of issue-specific frames in protest news coverage?

Research on Chile's Protest News Coverage
Protests and popular mobilizations have played an important role in Latin America's history (Johnston & Almeida, 2006), and have been studied from different perspectives and disciplines, including sociology (e.g., Somma & Bargsted, 2015), political science (e.g., Disi Pavlic, 2020), and mass communication (e.g., Valenzuela, 2013). In the case of Chile, the study of protests has gained more attention due to the many demonstrations the country has experienced in the last decade. Most of the recent mass communication literature on Chile's protests relates to social media use and protest participation (e.g., Valenzuela et al., 2012Valenzuela et al., , 2018, but framing research about protest coverage is still scarce, and mostly focused on issue-specific frames. For example, Pérez Arredondo (2016) found the news media systematically criminalized protesters and left aside their demands while covering Chile's college students' protests in 2011. Similarly, Sáez Gallardo (2019) found that news portrayal of Mapuche people (one of Chile's indigenous groups) depicted this group as extremely violent and deviant from the status quo. These findings indicate a pattern of negative coverage when it comes to protests and protest participants. However, these and other studies (e.g., Bonner & Dammert, 2021) about protest coverage in Chile are focused mostly on newspaper coverage. While the print press is said to be the main agenda setter worldwide (McCombs, 2014;McCombs & Ghanem, 2001), Chilean audiences are concentrated in other media, such as television, radio, and especially social media. In fact, the majority of Chileans rely on social platforms (particularly Facebook) to consume news, and the most trusted news outlet in the country is a radio news network, Radio Bío Bío (Newman et al., 2019(Newman et al., , 2020(Newman et al., , 2021(Newman et al., , 2022. And yet, radio news coverage remains understudied compared to print and television. Our study aims to fill this gap in the literature, at least to some extent, by looking at Radio Bío Bío's coverage of Chile's Estallido Social.

Chile's Political Unrest
While Chile has witnessed several protest movements in the last 10 years, the demonstrations linked to the social unrest in 2019 (popularly called Estallido Social) reached a breakpoint in the country's social and political context, with an unexpected degree of violence and aggressive reactions from the authorities and the police (Sehnbruch & Donoso, 2020). While some were surprised by this social eruption, there were clear cues announcing what was coming.
On October 7, 2019, groups of high-school students began small demonstrations in Santiago against increasing public transportation fares. While the protests started peacefully, the tensions between students and the police increased, and the chaos escalated quickly. By October 18, demonstrations had spread through the country, this time becoming a large movement of people marching the streets to protest the government as well as endless social injustice. However, messages in favor of dignity and equity were tainted by riots and violent acts performed by isolated groups. The images of buildings and public transportation in flames, and lootings in pharmacies and grocery stores, were the ones the news media broadcast to cover the protests, which ended up criminalizing part of the movement.
Chile's Estallido Social is part of a series of political crises in Latin America that began in 2019, demanding equity and dignity. Still, when the media covered these movements, the focus was on the chaos and violence, criminalizing the movements and exacerbating social conflict (Chacón & Rivera, 2020).

Methods
Data for this study come from stories published by Radio Bío Bío, the most popular radio news network in Chile, and the most trusted news outlet in the country, according to Reuters Institute (Newman et al., 2019(Newman et al., , 2020(Newman et al., , 2021(Newman et al., , 2022. As social media in general, and Facebook in particular, have become the main source for Chilean audiences to get informed (Newman et al., 2022), we content-analyzed news stories published on Radio Bío Bío's Facebook page. We produced two constructed weeks to cover two months of data-from October 18 (the day when the protests escalated) to December 18, 2019.
We built a Python script to retrieve all news stories published on Radio Bío Bío's Facebook page in those two constructed weeks. We retrieved around 2,500 stories and randomly selected 1,200 stories to code for 1 = unrest-related stories, and 0 = other stories. The material was coded by three undergraduate students unfamiliar with the study goals. Following Lacy et al.'s (2015) best practices for content analysis, the authors of this study developed the codebook and trained the coders, but did not code the material to avoid researcher bias. Coders were recruited from Pontificia Universidad Católica de Chile Summer Research Program, a university-wide initiative developed for undergraduate students to get involved in academic research. Based on a subsample of 120 stories (not included in the final sample), and using the ReCal3 software (Freelon, 2010), inter-coder reliability (ICR) reached 93% agreement and a Krippendorff's alpha of 0.84.
We identified 657 stories about the political unrest, from which we randomly selected a sample of 417 stories to code for generic and protest frames. We left the remaining 240 stories to train the coders and select a subsample of 40 stories to calculate ICR in this second stage. To identify generic frames, we coded for items validated by previous framing research (Burscher et al., 2014;Valenzuela et al., 2017) and described them in Table 1. As some stories aimed to provide the public with practical, useful information (e.g., subway stations not working), we also coded for public service (Berger & Milkman, 2012). Some stories did not fit into the five generic frames, so we created the option Other. To code for issue-specific frames, we used Hertog and McLeod's (2001) categories also extensively validated by research on protest frames (e.g., Harlow et al., 2020;. Our three coders reached ICR levels ranging from 91% of agreement (Krippendorff's alpha of 0.67) to 100% (Krippendorff's alpha of 1) for all the variables coded in the study. Table 1 describes measurements and  ICR values for each variable. To estimate the relationship between generic frames and protest frames, we modeled a multinomial logistic regression with protest frames as the multicategory dependent variable (with Does Not Apply as the reference category) and the five generic frames, public service, and Other as independent variables.

Results
Descriptive findings indicate news stories portrayed Chile's political unrest from either the Riot (26%) or the Confrontation (12%) perspectives, while a minority of stories paid attention to the protesters' motivations by using the Debate frame (7.5%). Just a few stories reported the unrest from the Spectacle frame (5%). However, half of the analyzed stories did not fit into any of the protest frames (see Figure 1), as an important proportion of the coverage focused on informing the audience about safe ways to move around the city. As such, those were unrest-related stories but did not necessarily address the unrest using protest frames.
In terms of generic frames, results show the most prominent frame was the attribution of responsibility frame (28%), followed by the conflict frame (27%). In other words, the coverage paid special attention to the conflict itself, and those responsible for it. Only 5.6% of the stories portrayed the unrest from the moral perspective, judging both the government and the protesters for hurting the country. The economic consequences (5%) and the human interest (3.5%) frames were at the bottom of the rank (see Figure 2). Similar to Figure 1, Figure 2 illustrates that stories tackled other perspectives, such as the public service perspective (providing practical information to audiences) and information that did not really fit into generic-frame categories.
RQ inquired about the nature of the relationship between generic frames and protest frames in the news coverage of Chile's political unrest. We modeled two multinomial logistic regressions to answer RQ. In the first model, we included Semetko and Valkenburg's (2000) five generic frames, as well as the public service and the Other variables. Due to the small number of stories presenting the economic consequences and the humaninterest frames (as illustrated in Figure 2), floating-point overflow occurred while computing the regression values for many cells (see Supplementary File). To reduce the number of missing values, we removed the economic consequences and the human-interest frames from the predictors and ran a second model (see Table 2), which allowed for a more parsimonious and accurate analysis of the data.
Results from the second multinomial logistic regression suggest that certain generic frames increase the presence of protest frames in the news coverage. The attribution of responsibility frame increased the presence of the Riot, the Confrontation, and the Spectacle perspectives, blaming protesters for the violent clashes, looting, and arson attacks occurring during the analyzed timeframe. In fact, most of the stories using both responsibility and Riot frames blamed protesters for violent episodes. The conflict frame was associated with a higher presence of both the Riot and the Confrontation perspectives, with stories emphasizing the mismatch between protesters and their grievances, and the official response from the government. Interestingly enough, the morality frame was the only generic frame positively associated with the Debate frame. Stories portraying the political unrest from a moral perspective relied on official and unofficial source quotes to suggest how the government should react and how protesters should behave, for both of them to reach at least some degree of understanding. At the same time, these stories were more likely to emphasize the motivations and reasons behind the unrest, therefore increasing the Debate perspective. Neither the public service variable nor the Other variable were significantly associated with the protest frames.

(a) Feelings
If the story employed adjectives or personal vignettes that generate feelings of empathy, rage, or concern, it was coded 1. If not, it was coded as 0.

(b) Emotional resources
If the story provided a human example or a human face to the issue, it was coded as 1. If not, it was coded as 0.

Morality frame (a) Moral message
If the story included a moral/ethical message or a social judgment, it was coded as 1. If not, it was coded as 0.
95% 0.72 Public service If the story provided the reader with useful information to make decisions, it was coded as 1.
If not, it was coded as 0.

Other
If the story was coded as 0 in all the previous categories, it was coded as 1 in this category.

95% 0.95
Protest frames If the story focused on the disruptive behavior of the demonstrators (and the effects the latter has on society), it was coded as Riot (1).
If the story focused on the confrontations between protesters and the police, it was coded as Confrontation (2).
If the story focused on the "spectacular" elements of the protest and its context in Chilean history (strangeness or unusualness, massiveness and scope, emotion, and drama), it was coded as Spectacle (3).
If the story focused on the protest agenda or demands, aiming to inform about the issues defended by the social movement, it was coded as Debate (4).
If the story did not cover protests/demonstrations, or if the protest was presented as a peaceful event, it was coded as Does Not Apply (5) 3.5% Figure 2. Distribution of stories according to generic frames. Note: Percentages add up to more than 100% as a story might present more than one frame. Notes: (a) Does Not Apply was used as the reference category in the dependent variable; (b) floating point overflow occurred while computing this statistic, and its value is therefore set to system missing; * p < 0.05, ** p < 0.01, *** p < 0.001.

Discussion
This study explored the relationship between generic frames and protest frames in the news media coverage of Chile's social and political unrest in 2019. By analyzing news stories published by Radio Bío Bío, our findings are consistent with previous research regarding protest coverage. The stories replicated the patterns that usually delegitimize the protest, as they focused on the violent acts and the depiction of protesters as deviant from the status quo (McCluskey et al., 2009;McLeod & Hertog, 1992). Our results show that frames related to protest delegitimization were prominent, with a higher presence of the Riot (26%) and Confrontation (12%) perspectives, similar to what  found. However, we also found a low presence of the Spectacle frame (5%), challenging what Harlow et al. (2020) found for Latin American countries. When looking at both sets of frames together, our results show consistent relation patterns. Given that Riot and Confrontation are the most-used protest frames in the coverage, it makes sense that the most prominent generic frame is the attribution of responsibil-ity. When covering the protest, the news media not only emphasized violent actions but also inquired about those responsible for such violent actions. Similarly, the presence of the Confrontation frame was linked to the clashes between protesters and the police portrayed in the stories, which in turn explains the presence of the conflict frame. By putting together generic and protest frames, we observe in detail the perspective used to depict the protest (a delegitimizing perspective) and the general portrayal used to build the news narrative (conflict and responsibility attribution), making evident what is important for the Chilean news media.
While the attribution of responsibility frame increased the presence of most of the protest frames, it did not increase the presence of the Debate frame. As such, the coverage did not focus on the social demands behind the protest or the course of action for the future. Consequently, the news media focused on describing and judging violent actions, but not necessarily on problem-solving. Harlow et al. (2020) found a high presence of the Debate frame, something we did not find in our study, as only 7.5% of the stories used this frame. Nevertheless, we want to point out the relevance of the Debate and morality frames correlation. Chile's social unrest originates from demands against social injustice, such as the right to access quality health and education services. In this case, the Debate frame questioned Chilean society as a whole from a moral perspective. Rather than presenting the protest in a positive way, it provided support to the protest demands.
These results highlight the nature of the relationship between both sets of frames, in which the presence of generic frames influences the salience of certain protest frames. For instance, when covering a riot, the most salient elements in the story aimed to identify the responsible for violent actions (when using the attribution of responsibility frame) or describe the sides in the conflict (when using the conflict frame). Consequently, the riot as an event was not necessarily the main focus of analysis, but certain riot features were determined by the most prominent generic frame used to portray the story.
Additionally, an intriguing finding was the high number of stories not fitting into any of the protest frames, as well as stories that did not use any of the generic frames. While issue-specific frames reveal what aspects are salient in the news coverage, it might be the case that Chile's Estallido Social does not necessarily fit the existing protest categories. More than "a protest," the Chilean case was a "protest process" where each day was different from the previous one, and many aspects of social life were affected by this process. Consequently, when covering protests as a socio-political process, the coverage focused on the riots, the debates, the spectacle, and the conflict, but also reported on other dimensions that are not necessarily part of the existent theory. Similarly, the "other" category in generic frames echoes Gronemeyer et al.'s (2020) call to include an "informative" generic frame, also based on the context of the study. As previously explained, the context where protest movements take place is important in explaining protest coverage (e.g., Kilgo, 2020;. Given the nature of the protest and its socio-historical context, the so-called Estallido Social might have challenged how the news media portrayed the event.

Conclusions
From our results, we were able to observe a direct relationship between generic frames and protest frames, in which the presence of the former seems to determine that of the latter. Protest frames focus on the presence of events and actions related to the demonstrations, while generic frames provide information about how the news media interpret and package the news. When analyzing both sets of frames together, the relationships between framing categories show that it is possible to study them from a mixed-frame approach, following Kozman's (2017) proposal.
Even though we observed elements of the traditional protest coverage (e.g., the exaltation of violence, protesters presented as deviants, etc.), it is important to acknowledge the nature of the protest, and the sociohistorical context might impact how the media portray the event. Moreover, it can transgress the traditional categories of framing by making salient aspects that are not considered in the original theories (as shown in the Other and Does Not Apply categories).
This study is not without limitations. First, our analysis relied on a single news provider, and as such, we cannot draw general conclusions about the country's media system as a whole. We analyzed news stories published on Radio Bío Bío's Facebook page because Radio Bío Bío is the most trusted news outlet in the country, and also because Chilean audiences rely mostly on social media (particularly Facebook) to get informed. While most framing research looking at Chilean protest has analyzed print news coverage, there is limited research looking at other types of news media that reach larger audiences. Thus, we looked at the Facebook page of a respected and highly consumed news source (Newman et al., 2019(Newman et al., , 2020(Newman et al., , 2021(Newman et al., , 2022). Yet, more research is needed to compare protest news coverage across print, radio, television, and online, to achieve a more holistic understanding of news framing when it comes to covering protests.
The findings of this study show that combining both generic and issue-specific frames is a helpful approach to understanding the complexities of protest news coverage. We believe this mixed-frame approach is useful to make sense of frame functions such as defining problems or finding solutions (Entman, 1993), and also to observe how the public perceives protest movements. For instance, news coverage of violent demonstrations might use the riot frame but emphasizing economic consequences versus attribution of responsibility might have different effects on the audience. While identifying such effects is beyond the scope of this article, future research could apply an experimental design to test this possibility.