The United States has experienced several high-profile mass school shootings over the past 50 years, but such events were not discussed as a social problem until the incident at Columbine High School in 1999 (Altheide, 2009; Elsass et al., 2014; Muschert, 2007; Muschert & Carr, 2006). Columbine was different from previous school shootings because it was covered by cable news media (Murray, 2017; Schildkraut & Muschert, 2014) and extensively by traditional print media (Chyi & McCombs, 2004). These types of events are often sensationalized within the media and produce a copious amount of media coverage (Muschert, 2007; Schildkraut et al., 2018). Accordingly, the media have become a significant avenue for school shootings research because consumers primarily experience these events through the coverage (Schildkraut et al., 2018). Within this line of work, scholars have examined the media coverage of prominent mass school shootings such as Columbine (Altheide, 2009; Chyi & McCombs, 2004; Larkin, 2009), Virginia Tech (Fox & Savage, 2009; Hawdon et al., 2014), and Sandy Hook (Murray, 2017; Schildkraut & Muschert, 2014). The findings from these studies suggest local media outlets closer to a mass school shooting tend to focus on the victims, shooter, and local community, while national media outlets tend to focus on macro issues related to shootings such as gun control.

The shooting at Marjory Stoneman Douglas High School in Parkland, Florida, on February 14, 2018, was widely covered by media outlets for an extended amount of time.¹ The Parkland shooting resulted in seventeen deaths and seventeen injuries. To date, only a few studies have examined media coverage and framing of this event, despite it being one of the deadliest high school shootings. Understanding how this shooting was covered is important considering most of the information the average person knows about mass shootings comes from the media (Schildkraut et al., 2018). In addition, media consumption of mass shooting news stories has been found to be positively associated with increased fear about experiencing this type of event (Burns & Crawford, 1999; Goode & Ben-Yehuda, 1994; Kupchik & Bracy, 2009; Levin & Wiest, 2018; Morrow et al., 2016), increases in gun legislation (Luca et al., 2020), and increases in gun ownership (Porfiri et al., 2019).

There is significant variation in the methods and focus of the few media studies that have focused on the Parkland shooting. For example, Aslett et al. (2020) conducted an analysis of tweets derived from both gun rights and gun control groups following the Parkland shooting to identify how each group respectively framed the problem. Rees et al. (2020) analyzed print media coverage to identify contributing factors to the Parkland shooting. Finally, Holody and Shaugnessy (2020) analyzed print media coverage of the Parkland shooting with a focus on comparing the salience and valence of fifteen frames across local and national outlets. While each contributes to the literature on the Parkland shooting, mass shootings, and media behavior, they do not account for how the media did or did not change their framing of the Parkland shooting in the weeks after it. That is, these studies of the Parkland shooting do not account for the dynamic process of framing (see Chyi & McCombs, 2004). Framing as a dynamic process suggests that the frames relied upon by the media in the immediate aftermath of the shooting may be different from those relied upon months after it has happened.

The current study extends previous research about media framing of school shootings to the Parkland incident by analyzing front-page coverage of the shooting published between February 15, 2018, and April 20, 2018, in The New York Times, USA Today, The Washington Post, Orlando Sentinel, Sun Sentinel and Tampa Bay Times. It utilizes methods established by Hawdon et al. (2014), who analyzed print media coverage of the Virginia Tech shooting through Pennebaker and Harber’s (1993) Social Stage Model of Coping.² We extend this and prior work by examining how the distance between a print media organization and the Parkland community influenced the framing of the event across three time periods. Finally, by applying the Social Coping Model to a shooting beyond the Virginia Tech shooting, we attempt to provide a reliable coding schema that fosters cross-event generalizability and comparison.

Literature Review

Media and Mass Shootings

While definitions vary, many scholars define mass shootings as incidents that result in the deaths of four or more people by gunfire, excluding the offender (Duwe, 2004; Fox et al., 2020; Krouse & Richardson, 2015; Lankford & Tomek, 2018; Meloy et al., 2001; Peterson & Densley, 2019). Despite variation in how mass shootings are defined, most individuals indirectly experience mass or school shooting events through the media. For example, in the year after the Columbine High School shooting, the three major news networks (ABC, CBS, NBC) crafted over 300 stories about the shooting (Robinson, 2011). Thus, analyses of media coverage about mass shootings are instrumental for the development of knowledge about incidents themselves, media behavior, and the impacts of such events.

Researchers have concluded mass shootings are portrayed in the media for an extensive amount of time and not all covered in the same manner. For example, in an analysis of New York Times articles from 2000 to 2012, Schildkraut et al. (2018) found that race/ethnicity, specifically the shooter being of Asian and other descent, and the number of victims were the most important factors in portraying these events in the media. Similarly, an analysis of television coverage of mass shootings from the three major television networks (ABC, NBC, and CBS) over a twenty-five year span (1989-2014) found that mass shootings with ten or more fatalities receive more coverage than shootings with six to nine fatalities, which in turn receive more coverage than when there are less than six fatalities (Luca et al., 2020). Another important factor is the location of the shooting. Here, research has found that shootings taking place at government buildings (Fox et al., 2020; Silva & Capellan, 2019) or in schools (Fox et al., 2020; Schildkraut et al., 2018; Silva & Capellan, 2019) often take precedence in coverage over shootings that occur in other locations.

Scholars have also been able to tie media coverage of mass shootings to specific social consequences. First, increased media attention on these events has been linked to fear of such incidents. In a survey of 212 adults, Kupchik and Bracy (2009) found that those fearful of mass shooting victimization were more interested in reading a story about a mass shooting. Likewise, an analysis of school crime and violence news published by USA Today and The New York Times found that these sources persistently remind readers about the potential of school violence and frame school crime as getting worse (Levin & Wiest, 2018). Media outlets often reference the last extreme school shooting to emphasize the devastation of the latest attack (Chyi & McCombs, 2004; Schildkraut, 2016). This distorts the audience’s views of the actual event and can lead to an audience developing more punitive attitudes or a general sense within the public that the number of shootings increased (Schildkraut et al., 2015). Thus, media framing can continue to generate fear of school shootings, despite evidence that the number of events has have declined (Haan & Mays, 2013).

Finally, research has evidenced the impact of mass shooting coverage on future shooters, the dissemination of mass shooting information, and gun related issues. In an ethnographic analysis of the media coverage involving three school shootings (Columbine, Virginia Tech, and Sandy Hook), Murray (2017) suggests stories of shootings can enable future shooters by providing copious amounts of information to the public about the offenders.³ Similarly, Sumiala and Tikka’s (2011) examination of YouTube videos about four mass school shootings (Columbine, Virginia Tech, Jokela, and Kauhajoki) found that YouTube facilitated the circulation of videos about school shootings including videos produced by shooters themselves. This circulation promotes violent social imaginaries and can blur what is considered professional news versus non-professional news content (Sumiala & Tikka, 2011). Finally, a recent study of mass shootings over a twenty-five year period finds that a single mass shooting leads to a 15% increase in the number of firearm bills introduced within a state in the year after an incident and is correlated with increased television news coverage (Luca et al. 2020). To complement this finding, an analysis of tweets derived from gun rights groups and gun control groups following the Parkland shooting found each group used Twitter to advance their respective policy narratives (Aslett et al., 2020).

Media Framing and School Shootings

Framing is an important tool used by the media when presenting information (Chermak, 1994; Gerbner & Gross, 1976; Surette, 2015). Media outlets frame information with a specific narrative and reframe stories to recycle them. Entman (1993) argued that framing was the process of selecting specific aspects of a “perceived reality” and making them prominent to define a problem, discuss the causes, make moral judgments, and possibly finding a solution. Scheufele and Tewsbury (2007) wrote that framings are “invaluable tools for presenting relatively complex issues…efficiently and in a way that makes them accessible to lay audiences because they play to existing cognitive schemas” (p. 12). Thus, how the media chooses to frame news is significant as it can affect consumers in a variety of ways, including gaining public support on an issue or the furthering of political arguments (Chyi & McCombs, 2004; McCombs & Shaw, 1972).

Scholars have specifically focused on media framing of mass school shootings. This interest increased dramatically after the highly publicized shooting at Columbine High School in 1999 (Murray, 2017; Muschert & Carr, 2006; Schildkraut & Muschert, 2014). School shootings represent a unique form of extreme violence because of their location and the greater likelihood that youth are victimized by these types of incidents. Since Columbine, analyses of media framings about school shootings have been instrumental for the development of knowledge about these incidents, media behavior, and the impacts of such events. For example, analyses of newspaper coverage of Columbine and nine other shootings concluded that coverage of these events were focused on individuals immediately after their occurrence, but later, these events were discussed in a larger social context that evoked conversations about issues, such as gun control (Chyi & McCombs, 2004; Muschert & Carr, 2006). In addition, a comparative content analysis found that coverage of the Sandy Hook shooting more often focused on the actions of the educators involved rather than focusing on the shooter when compared to the Columbine shooting (Schildkraut & Muschert, 2014). Finally, Frymer’s (2009) examination of print and television news media surrounding the Columbine shootings concluded that the media perpetuated a narrative of “youth violence and alienation” (p. 1390). Recently, Rees et al. (2020) conducted a root cause analysis involving 282 articles from ten online new sources to identify contributing factors to the Parkland shooting. They found that factors fit within four themes: policy (gun legislation), person (mental illness), environmental (culture), and equipment (large ammunition).

Much of the early research on media framing of school shootings only examined coverage from The New York Times (see Chyi & McCombs, 2004; Muschert & Carr, 2006; Schildkraut & Muschert, 2014). These studies brought forth valuable insights, but they were unable to examine differences in framing across media outlets. Holody and Shaugnessy (2020) addressed this gap by comparing local and national print media coverage of the Parkland shooting and focusing on comparing the salience and valence of fifteen frames. The study found that both local and national print media framed the shooting negatively (valence), while mostly framing their coverage of the shooting around issues of gun control (salience). In another study, Hawdon et al. (2014) provided a content analysis of 854 news stories about the Virginia Tech shooting collected from The Washington Post, Chicago Tribune, Los Angeles Times, Wall Street Journal, and Roanoke Times. The findings suggested media location and time affected which frame was used. Specifically, the location of the news source impacted which aspects would be accented in the presentation of the story. Importantly, the study found that the focus of the coverage would change in the weeks after the shooting, and how the focus changed depended on the media location. To summarize, media outlets often frame mass shootings utilizing a variety of techniques. Further, evidence suggests that over time, both the framing of a shooting and the amount of coverage of a shooting garners can shift (Chyi & McCombs, 2004; Hawdon et al., 2014; Holody, 2020; McQuail, 2010).

Theoretical Framework/Foundations

The current study follows in the logic established by Chyi and McCombs (2004) concerning the study of framing “as a process over time” (p. 26). They offered a pathway for cross-event generalizability for framing research through a schema that took into consideration “space” and “time.” Space here refers to categorizing frames that emerge from event coverage by their socially distant qualities. These frames refer to how the event is presented in the media by focusing on individual victims (individual frame), the community where the event occurred (community level frame), the larger geographic area surrounding the event (regional level frame), society or national significance (societal frame), and international significance (international frame). Time refers to categorizing frames by whether event coverage invokes either the past, present, or future. They then applied this framing schema through a content analysis of 170 New York Times articles covering the Columbine shooting. With respects to space, the study found that most of the coverage evoked the societal frame by discussing the shooting through the larger social context and focusing on issues like gun control. Finally, the study found that coverage tended to focus on the present with respect to time.

The Chyi and McCombs (2004) study offered a framing schema that could be used to understand media framing across different school shootings. Importantly, the study also evidenced that framing was a dynamic process, that frames were subject to change throughout the course of an events lifespan. That is, they found that even while the coverage favored the societal frame, the majority of coverage that evoked the societal frame came later in the month following the event. While this was a significant finding, the study was limited to one media outlet (The New York Times) and an analysis of media coverage spanning a month. This left room to expand on framing as a dynamic process. In addition, the study’s use of space and time is limited to how they are evoked within coverage. That is, news coverage of a shooting can discuss societal issues surrounding these incidents (space) while also discussing them with reference to the past, present, or future (time). We suggest an additional approach to understanding the significance of space and time towards the dynamic process of framing is to address how physical space, and time influence the dynamic process of framing. That is, if frames can change in the aftermath of a school shooting, how does a media outlet’s location intersect with time since the shooting to influence how it is framed?

Hawdon et al.’s (2014) analysis of media coverage about the Virginia Tech shooting answered these questions by applying the Social Coping Model, which emphasizes the coping and healing of the collective after a traumatic event and how this process plays out in stages across time. Using survey data following the Loma Prieta Earthquake and the start of the Persian Gulf War, Pennebaker and Harber (1993) found that individuals would talk and think about these events the most in the first few weeks following the event, but discussions and thoughts would return to normal by six to twelve weeks. Based on the results, Pennebaker and Harber (1993) suggested that individuals proceed through three phases while coping with the aftermath of an event. The Emergency Phase occurs two to three weeks after an event when most discussion occurs because the event is fresh in the minds of the public. Next, the Inhibition Phase occurs three to six weeks after an event, when members of the society no longer openly discuss the event but continue to think about it. Finally, the Adaptation Phase occurs when members of the society no longer discuss or think about the event.

We argue that, after a high-profile school shooting, the media may simply be framing the various stages of collective coping as outlined by the Social Coping Model. Despite its relevance, the only study to apply the Social Coping Model to print media coverage of a mass school shooting is Hawdon et al. (2014). That study proposed that the local media should be printing more stories about the Virginia Tech shooting due to proximity, but that as a result of competition and relationship building, they would likely avoid discussion of broader issues (i.e., gun control), and focus on the victims and the community in the immediate aftermath of a mass shooting (i.e., the Emergency Phase). They found that most of the sampled articles were published during the Emergency Phase, with publication declining throughout the following phases as predicted by the Social Coping Model. In addition, while overall coverage of the Virginia Tech shooting declined, local papers were found to be more likely to continue publishing articles during the Inhibition and Adaptation Phases than national newspapers. The cross-event generalizability of the Social Coping Model remains unknown given that the study only focused on one shooting. The current study addresses this gap by applying the Social Coping Model to the Parkland school shooting, which allows for the possibility of comparing the dynamic framing of two high-profile shootings.

Based off previous literature (Chyi & McCombs, 2004; Hawdon et al. 2014; Muschert & Carr, 2006; Schildkraut & Muschert, 2014), the particular saliency of both location and time in influencing the media’s framing is a principal interest of the current study. Specifically, the Social Coping Model, as applied by Hawdon et al. (2014), is primarily used to guide this study. In line with this framework, we seek to answer the following research questions and test the associated hypotheses:

RQ1: What is the media location’s impact on the amount of coverage of the Parkland shooting across the three stages of the Social Coping Model?

H1: Local papers will print more stories regarding the shooting than national papers.

H2: Local papers will be more likely to report on the shooting during the Adaptation Phase.

RQ2: What is the media location’s impact on the framing of the Parkland shooting across the three stages of the Social Coping Model?

H3: During the Emergency Phase, local papers will be more likely to focus on the shooter.

H4: During the Emergency Phase, local papers will be more likely to focus on victims.

H5: During the Emergency and Inhibition Phases, national papers will be more likely to focus on broader issues.

H6: During the Adaptation phase, the number of articles focusing on new information about the shooting will differ by location of the media.

RQ3: What is the media location’s impact on framing of the Parkland community across the three stages of the Social Coping Model?

H7: During the Emergency Phase, local papers will be more likely to depict the afflicted community as experiencing collective trauma and grief.

H8: During the Emergency Phase, local papers will be more likely to depict community solidarity.

H9: During the Emergency Phase, local papers will be less likely to focus on conflict in the local community.

Methodology

Data

The current study used a purposive sampling strategy to collect and analyze a sample of 325 articles from six newspapers about the Parkland shooting published between February 15, 2018 and April 20, 2018. The start date of February 15th was selected because it was the day after the shooting, while April 20th marked the end of the three-week period covered by the Adaptation Phase. The six newspapers included three local newspapers (Tampa Bay Times, Sun Sentinel, and Orlando Sentinel) and three national newspapers (USA Today, The New York Times, and The Washington Post). Local newspapers were purposefully chosen based on their location in relation to Broward County and readership. The Tampa Bay Times (2019) was considered Florida’s most circulated newspaper, followed by the Orlando Sentinel (Agility PR Solutions, 2019). The Sun Sentinel is in Broward County, where Parkland is located. An equal set of national newspapers were selected for overlap with Hawdon et al. (2014) and readership. At the time of the analysis, USA Today ranked number one in terms of readership, followed by The New York Times (Misachi, 2017). While The Washington Post ranked seventh, it was included for its significance in the Virginia Tech shooting (see Hawdon et al., 2014).

National and local outlets were selected to analyze geographic differences in media framing. Articles that met the inclusion criteria of pertaining to the Parkland shooting, presenting the victims of the shooting, discussing the community, or presenting news concerning broader issues (i.e., gun control policies or mental health) were retained for subsequent analysis. LexisNexis and NewsBank were used to search for articles pertaining to the shooting. The search terms included: Parkland, Parkland school shooting, and Marjory Stoneman Douglas High School. The results were narrowed by timeline (February 15, 2018 – April 20, 2018), publication type (newspapers), location (Parkland, Florida), and news company.

A deductive or a priori coding strategy was used. This procedure relied on the determination of field codes of those used by Hawdon et al. (2014). This allowed for appropriate comparison between two separate mass shootings while also providing the chance to ensure reliability and validity of the coding scheme developed by Hawdon et al. (2014). Each article was coded based upon 1) focus (i.e., the overall frame such as articles focusing on the shooter or victims); 2) reporting signs of community solidarity; 3) reporting of division or conflicts occurring after the shooting; and 4) whether the Parkland community was depicted as damaged or collectively grieving. The current study investigates each of these frames by the geographic coverage of media outlets across time. Specifically, it investigates the frames of the sampled articles according to the three phases defined by the Social Coping Model. Phase 1, the Emergency Phase, covered the first three weeks after the shooting (February 15 through March 7, 2018). Phase 2, the Inhibition Phase, covered the three weeks after Phase 1 (March 8 through March 29, 2018). Finally Phase 3, the Adaptation Phase, covered the three weeks after Phase 2 (March 30 through April 20, 2018).

Variables

Table 1 displays the conceptualization and operationalization of our frames. Frames were housed within two overarching categories: focus and community impact. Our coding schema (focus, community impact, and location) draws from Hawdon et al. (2014) to provide cross-event generalizability of the coding schema and comparison of media framings across two separate shootings (i.e., framing of Virginia Tech vs. Parkland). The mutually exclusive coding of the news articles’ dominant frame was based on their focus on a) the shooter, b) victims, c) broader issues, and d) news. For example, an article that discussed the shooter and his background in addition to information about the victims (those who were killed and those who were involved but not injured) were coded as shooter-focused because the article began and ended with details about him (McMahon et al., 2018). The next category concentrated on the impact of the Parkland shooting on the community. Here articles were coded for a) community solidarity, b) divided and conflicted, and c) grieving and shattered. Articles were coded as reporting signs of community solidarity, division or conflict, grief, or not reporting these signs. As such, these categories are not mutually exclusive, and articles could present signs of community solidarity in addition to signs of grief and/or conflict. Finally, newspapers’ media location was coded as local or national for most of the analyses in this study. The Tampa Bay Times, The Sun Sentinel and The Orlando Sentinel were considered “local” papers, while The New York Times, The Washington Post, and USA Today were coded as “national” papers.

To establish inter-coder reliability, two coders coded 65 of the 325 articles (20%) and Cohen’s kappa scores were generated. Scores were .713 for community damaged/grieving, .742 for community solidarity, and .732 for community division/conflict. Scores reflected substantial agreement above established threshold (see Gottschalk, 2014). An additional round of discussions was conducted between the two coders to increase agreement before one coder then coded all 325 articles.