Abstract
Many studies treat news framing as a static phenomenon and fail to capture the framing process during breaking news events. This study explores the dynamic nature of news framing through an analysis of The New York Times’ continuous reporting on the September 11 attacks. The findings reveal four key characteristics of dynamic framing: stabilization, re-focalization, frame shifts, and frame convergence. Frames evolved from exploratory at the initial stage to increasingly focalized and stabilized in later stages. Re-focalization occurred as primary natural frames lost dominance to extended and marginal frames. Frame shifts, including frame desertion, transformation, and absorption, highlighted changes in news framing as reporting developed. Frame convergence marked the final phase, with dominant frames integrating others to create cohesive narratives. This study provides a model of news framing as a dynamic focalizing process, offering insights into the evolution of news discourse during continuous reporting. The findings contribute to framing theory by emphasizing the fluidity of news framing in breaking news contexts and hold implications for understanding framing in other types of news events.
Introduction
Since Bateson (1955, pp. 39–51) proposed “frame” and after it was elaborated by Goffman (1974), it has become a concept widely used in communication studies, sociology and linguistics (Yuan & Wang, 2017). Frames are considered to be the socially shared organizing principles to construct reality (Goffman, 1974). To frame is to select, re-arrange, and make some aspects of reality more salient (Entman, 1993; Gamson & Modigliani, 1989). As D’Angelo and Kuypers put it, “hardly an issue of a communication journal is published today without a framing study” (2010, p. 1).
News making is characterized by the journalistic process of selecting and restructuring of news materials (Tuchman, 1978; van Dijk, 1988). In the news making process, newsmakers use frames to give meaning to news stories (Valkenburg et al., 1999). “News is easily the most prominent discursive site in which communication researchers strive to understand what framing is and how framing works” (D’Angelo & Kuypers, 2010, p. 1). In news framing studies, researchers have discussed the concepts of framing (Entman, 1993) and the relationship between framing and other concepts like agenda setting (McCombs & Ghanem, 2001). Besides, the production of news frames in the newsroom (Gitlin, 1980), frames in the media text (Gamson & Modigliani, 1989; Ojha, 2010), and the framing effects (Brewer & Gross, 2005) have been examined as well.
Current scholarship predominantly adopts static approaches that examine: (a) media framing of an issue like climate change, immigration, to uncover media manipulation (e.g. Huang & Liu, 2025; Simonsen, 2024); (b) in a breaking news event, how a specific topic was presented at a fixed time (e.g. Imtihani & Mariko, 2013). Few studies explored how a breaking news event was framed when it was continuously reported. In the continuous reporting of a breaking news event, there will be a formative process for the journalists to frame the event as the perceptions of the event and news values keep changing. The framing of news discourse will change until it reaches a stage where it finally gets normalized. Most previous studies focus on this normalized stage of news framing by examining reports of news events whose norms have normalized as a result of similar occurrences in the past. Framing in this sense, seems static if the formative process is left out.
The rationale for this study lies in understanding how journalists construct and adjust interpretive frames as news events unfold. Framing enables journalists to organize complex information and make sense of emerging crises by highlighting certain causes, actors, and consequences while downplaying others. In breaking news events such as 9/11, frames serve both cognitive and communicative functions: they help reporters quickly interpret uncertain realities and guide audiences toward shared understandings of legitimacy and responsibility. By tracing how dominant and marginal frames evolve over time, this study provides insights into how journalistic framing mediates the process through which the public makes sense of rapidly changing situations.
We will first review the typologies of frames and studies on news framing, and then present the theoretical and methodological consideration of the study; after that, we will elaborate dynamics of news framing in terms of stabilization of the categories of frames used in three phases; stabilization of the categories of frames used each day, re-focalization of frames, shift of frames and convergence of frames; finally, we will summarize our findings and propose contributions of this study.
Literature Review
Typologies of Frames
Understanding the various typologies of frames is essential for comprehending how media shapes public discourse and opinion.
Master Frames and Subframes: Master Frames are macro-level interpretive structures that shape collective understanding and action in social movements (Snow & Benford, 1992). They serve diagnostic (problem identification), prognostic (solution formulation), and motivational (action mobilization) functions, providing broad, adaptable narratives. Subframes are movement-specific adaptations of Master Frames, refining or extending their logic to address particular issues. For example, the feminist movement’s “gender equality” subframe stems from the civil rights Master Frame. While Master Frames provide foundational structures, Subframes contextualize and implement their core logic.
Issue-Specific Frames and Generic Frames: de Vreese (2005) differentiates between issue-specific and generic frames. Issue-specific frames are tailored to particular topics or events, shaping interpretations within a specific context. In contrast, generic frames—such as conflict, human interest, and economic consequences—apply across various issues, providing a consistent structure for news presentation. This distinction highlights the adaptability of generic frames and the contextual depth offered by issue-specific frames.
Episodic and Thematic Frames: Iyengar (1991) distinguished between episodic and thematic frames. Episodic frames highlight specific events or individual cases, emphasizing personal responsibility while overlooking structural factors, leading audiences to blame individuals rather than recognizing systemic causes. In contrast, thematic frames provide broader context by addressing institutional factors, encouraging viewers to consider societal responsibility. For example, reporting on poverty through an episodic frame may focus on an individual’s struggle, whereas a thematic frame examines economic policies and structural inequalities.
The commonality across different classifications is the distinction between broad and specific context frames. Broad frames (such as Master Frames and Generic Frames) are generalizable and applicable across a wide range of issues. They provide interpretive structures that focus on macro-level themes—such as social, or political issues. Specific context frames (such as Subframes and Issue-Specific Frames), are more narrowly focused, providing context-specific interpretations that emphasize particular details, backgrounds, and in-depth analysis.
Studies on News Framing
Studies on news framing include theoretical discussions, methodological approaches, and empirical studies.
Theoretical discussions on news framing center on three key areas. First is the definition of frames. Gitlin (1980) sees them as cognitive and organizing patterns in discourse; Gamson and Modigliani (1989) call them central organizing ideas; Entman (1993) adds that frames define problems, assign causes, make judgments, and suggest remedies. Second is the link between framing and agenda setting. McCombs and Ghanem (2001) differentiate them as “what to report” versus “how to report,” while Scheufele (2000) notes that framing depends on the frame’s relevance rather than its repetition. The third area concerns the relationship between news value and framing. Boesman and Van Gorp (2018) argue that frames arise from cultural concepts, while news values emerge from journalistic routines, guiding story angles and shaping media narratives.
Research methods for studying news framing include social scientific, rhetorical, and semantic network analysis. The social scientific method uses quantitative analysis. Gamson and Kathryn (1983) introduced interpretive packages. van Gorp (2010) recommends combining inductive and deductive methods to reduce subjectivity. Rhetorical approaches, like Kuypers (2010), focus on how frames shape meaning, often critiquing media bias through themes and norms. The semantic network approach, proposed by Baden (2018), treats news texts as interconnected nodes, allowing researchers to trace how frames evolve across media over time.
Empirical studies on news framing examine how media construct frames around issues or events and their impact on public perception. These studies can be grouped into three categories: framing of issues, framing of events, and framing effects. Studies on framing events, such as Chyi and McCombs (2004), focus on how frames evolve over time. They analyze the framing of the Columbine school shootings, observing shifts from individual-level frames to societal-level frames. Ojha (2010) compares media coverage of the 2008 Mumbai terrorist attacks, finding that national and international newspapers used different frames, with Indian press emphasizing domestic politics and British press framing it as a potential conflict between India and Pakistan. These studies highlight how journalists adapt frames to maintain the relevance of news stories and how different media outlets shape the narrative through ideological lenses.
However, many studies treat news framing as a static phenomenon and fail to capture the framing process during breaking news events, where routines and perceptions may influence framing development. Although news events selected in some studies were breaking news, for that September 11 attack provided a background for the perception of Mumbai terrorist attack in 2008, and the coverage of gun shooting was also familiar to the journalists, their routines were not completely interrupted to observe a relative entire framing process.
Methodology
News Framing: Conceptual Foundations
News making is not a passive reflection of reality but an active construction process in which journalists organize daily events into meaningful narratives (Tuchman, 1980). As Goffman (1974) argued, frames function as cognitive structures that enable individuals to interpret fragmented events by connecting them to culturally embedded schemas. Gamson and Modigliani (1989) further defined frames as central organizing ideas that structure the interpretation of news content.
Frames do more than organize information; they shape perception. Gitlin (1980) emphasized that frames are recurrent interpretive patterns that highlight certain angles while downplaying others, thereby influencing which issues gain prominence in public discourse. Through this selective emphasis, frames guide audience attention, affecting attitudes and behaviors. Entman (1993) conceptualized framing as a process that defines problems, diagnoses causes, makes moral evaluations, and suggests remedies through the activation of culturally resonant meanings. Thus, news framing operates simultaneously as a cognitive mechanism and a communicative strategy. It enables journalists to interpret complex events and, at the same time, influences how audiences understand and emotionally respond to those events (Entman, 1993; Gitlin, 1980). While framing can facilitate public understanding and collective action, it also raises ethical concerns when used to legitimize specific ideological positions or mobilize support for contested policies.
Recognizing this duality, the present study adopts a dialectical perspective that acknowledges both the constructive function and persuasive power of framing. It defines news framing as the process through which a central organizing idea is used to structure news discourse and as the dynamic evolution of this organizing idea over time. In this sense, framing is not static but develops across phases of news coverage, aligning with emerging interpretations, shifting emphases, and broader socio-political contexts
Data Selection
In this part, we will introduce criteria of the news event selection and our data collection.
Special Criteria of the News Event Selection
This study investigates the dynamics of news framing. First, it explores how journalists frame a rarely reported, unexpected breaking news event—an event unfamiliar to both media and audience that disrupts standard perceptions, routines, and journalistic habitus, prompting tentative framing. Second, it examines whether a frame shift occurs from breaking news to more routine reporting. This requires a news event of sufficient duration to provide ample reports. Third, the study tracks how frames evolve over time, which calls for daily reporting to trace these shifts.
Therefore, the news event has to meet three requirements. Firstly, it has to be a breaking news event which has rarely been reported, so that the normal perceptions of the event and norms of how it should be reported are interrupted, not like a planned conference, of which there are already normal perceptions and routines. Secondly, the length of the coverage has to amount to a certain period so that frame shift can be observed. Thirdly, there should be reports released each day to figure out the dynamics of framing process.
Selection of News Event
The September 11 attacks are chosen as the news event because they meet key criteria: suddenness, unpredictability, limited prior coverage, and a transition from breaking to routine reporting—ideal for examining framing dynamics. First, the crashes into the World Trade Center towers and the Pentagon created confusion over the nature, causes, and implications of the event, disrupting normal journalistic routines. Second, 24 days of continuous coverage allow tracking frame evolution. Third, a substantial corpus of over 800 reports, averaging 30 per day, offers a robust data base. Therefore, although (9/11 news reporting dates back to the past decade, and there are extensive prior researches on the social, political, and media impact of 9/11, the event remains a strong case for studying frame dynamics in news discourse. Its treatment provides value to the research project.
Data Collection and Research Questions
The New York Times is selected for its global prestige, large circulation, and Pulitzer Prize-winning 9/11 coverage in 2002. Data collection follows three steps. First, we consulted paper editions from September 12, 2001 to identify keywords and the data range, settling on “attack,”“911,” and “terror” as search terms, with October 8, 2001 as the cutoff—marking a shift in focus to the war in Afghanistan. Second, articles were retrieved from LexisNexis using the keywords and filtering by date and source. Third, the electronic corpus was cross-checked against the print edition to identify missing reports; ten articles were added using headline searches. The New York Times divides its 9/11 coverage into three labeled phases: “A Day of Terror” (Sept 12), “After the Attacks” (Sept 13–17), and “A Nation Challenged” (Sept 18–Oct 8). Table 1 provides an overview of the data. 812 articles over 24 days comprise the corpus.
An Overview of the Data.
The research question that guides our exploration is:
Analytical Procedure
Recent studies have increasingly emphasized the temporal and dynamic nature of media framing (Jiang et al., 2016; Klein & Amis, 2021). Rather than treating frames as static entities, dynamic framing approaches trace how interpretive patterns evolve across different stages of news coverage. In parallel, computational methods such as dynamic topic modeling (Blei & Lafferty, 2006) and longitudinal frame analysis (Lecheler & de Vreese, 2016) have provided tools to quantify such transformations over time. While these models capture large-scale trends, they often overlook the discursive and cultural logic of frame construction. Therefore, this study adopts a qualitative–quantitative mixed method that integrates corpus-based tracking with contextual interpretation to examine both the temporal variation and semantic coherence of frames.
To further ground the analysis in empirical observation, the study followed key tenets of Grounded Theory (Glaser & Strauss, 1998) in the process of frame identification and categorization. Specifically, the analysis proceeded through iterative stages of (a) coding the framing and reasoning devices of each news text; (b) grouping these codes around emerging “axes of meaning” that captured shared interpretive orientations; and (c) naming the resulting frame categories based on their dominant semantic and rhetorical features. This inductive process allowed frames to emerge directly from the data rather than being imposed a priori, ensuring that the evolving structure of frames reflects both the discursive logic of news construction and the dynamic shifts in meaning over time.
Identifying and Analyzing Frames
To identify hidden frames in news texts, this study uses content analysis, following van Gorp’s (2010) methodological guidelines. The full corpus consists of 812 articles published between September 12 and October 8, 2001. From this corpus, a stratified temporal sample of 140 articles was used to identify the frame categories d. After the frame typology was established, these frames were systematically applied to the entire corpus of 812 articles for quantitative coding and trend analysis.
Identifying Frames
This study follows three steps: identifying framing and reasoning devices, organizing them into a matrix with frames as rows and devices as columns, and then extracting dominant frames.
Step 1: Code the Framing and Reasoning Devices of the News Text
Coding focuses not on content but on how the text is framed. The aim is to identify textual indicators of frames. Each news article serves as the analytic unit. Both coders examine framing and reasoning devices. Framing devices includes metaphors, lexical choices, depictions, and historical references. Reasoning devices are the implicit logic connecting these elements (van Gorp, 2007). Since headlines, leads, and opening paragraphs often present the macro-proposition and set the interpretive frame (van Dijk, 1988), special attention is paid to these sections.
Lexical choices are identified through a combination of corpus analysis and manual coding. Keyword lists are generated using AntConc (Anthony, 2014). The British National Corpus (news subcorpus) serves as the reference corpus. For instance, in the article “A Somber Bush Says Terrorism Cannot Prevail” (Sept 12, 2001), keywords like “president,”“attack,”“nation,” and “security” were extracted.
On the other hand, many depictions and phrases can’t be auto-detected, so coders manually review the headlines and opening paragraphs. Table 2 shows the framing devices and reasoning devices in an excerpt.
Framing and Reasoning Devices in a Sampled News Text.
Step 2: Arrange the Framing Devices and Reasoning Devices Around “Axes” of Meaning
This step aims to analyze the framing devices and reasoning devices to get the possible frames which lie at the core of the package. Comparisons will be made to find the similarities and differences of the devices found in step 1 to reduce dimensions. A frame matrix will be built by filling in the rows with the summarized frames, and columns with the framing and reasoning devices. Table 3 is the frame matrix constructed out of Table 2.
Frame Matrix of an Excerpt.
Step 3: Name the Frames
The third step aims to construct frames. Since this study takes each news article as the analytic unit, there might be more than one frames identified. Comparisons again have to be made among the identified frames to further reduce and merge those frames. If there are still more than one frames, our solution would be to resort to both the news article to see which one is the dominant one, and to the headlines to see which frame is the one that headline is most closely related to. Take the excerpt in Table 2 as an example, the four frames in the frame matrix, that is, “Anti-terrorism war,”“Use force as a defensive war,”“Justice of war,”“Supported war,” can be further reduced and summarized as “Anti-terrorism.”
When naming the frames, we again follow van Gorp’s (2010) guideline that when the journalist explicitly mentions the frame, we will adopt the journalists’ naming; when the frame is not explicitly mentioned, we would make an association with the culture motives and consult relative literature to name the frame.
Step 4: Evaluate the Classification of Frames
Three criteria will be followed:
Thickness of the frame: the frame can be fully described by a list of framing and listing devices.
Abstraction of the frame: the frame is abstract enough and the chain of reasoning devices is complete and logically consistent.
Applicability to define other issues: the frame is not limited to a specific issue so as to beyond the scope of specific research topic and applicable to define other issues (van Gorp, 2010, pp. 97–99).
Two coders participated in the framing analysis: the author and one trained member of the research team. Both coders received prior training in identifying framing and reasoning devices following van Gorp’s (2010) framework. To ensure reliability, both coders independently coded a randomly selected 20% sample of the news corpus. Inter-coder reliability was calculated using Cohen’s κ = 0.84 and Krippendorff’s α = .81, indicating substantial agreement. Simple percentage agreement reached 89%. Disagreements were resolved through discussion and consensus, and the reconciled codebook was then applied to the remaining corpus. The final codebook, including operational definitions and illustrative examples, is provided in Table A1.
Results
Categories of Frames
As is shown in Table 4, 31 frames have been identified. Detailed elaborations of each frame, along with examples, can be found in Supplementary Material S1. The five most frequently used frame in the descending order are “Anti-terrorism,”“Disastrous Influence,”“Crisis Alleviation,”“Conspiracy,”“Threatened Security.” The event is mainly framed from an “Anti-terrorism” perspective. The social, economic damage caused by the attack, emergent measures as alleviation, the damage, interrupted domestic life, threatened security, have been also the dominant organizing ideas.
Categories of Frames.
In the subsequent analysis, the evolution of news frames was traced according to these operational criteria to identify patterns of dominance, desertion, and stabilization over time. The stabilization of frame categories was identified when the number of distinct frame categories used per day declined and then remained constant for three consecutive days, indicating that no new categories were introduced and no existing categories were dropped. A frame was considered dominant when it appeared in reports every day during a given stage with a daily frequency of two or more. Conversely, a frame was classified as deserted when it no longer appeared in any subsequent reports.
Stabilization of the Categories of Frames Used in Three Phases
Stabilization of frame categories means that no frames were abandoned or newly introduced. This was indicated by the decrease of the categories of frames after a trial and desertion from September 12 to 16, and by the decrease of the number of newly appeared frames from September 12 to 17, with no new frames appeared after September 17.
Decrease of the Total Categories of Frames
A trial-and-error process in the use of frames was observed from September 12 to 17. Some frames were initially used as expedient perspectives to organize the news story but were later abandoned. Between the first and second phases, following the initial day of the attack, 10 frames were deserted. Similarly, between the second and third phases, four additional frames were abandoned. The detailed information of these deserted frames is presented in Table B1.
Fourteen frames were tentatively used and abandoned before September 18. Figure 1 illustrates that the rate of frame desertion slowed from September 13 to 18, and after September 18, no frames were abandoned. This suggests an exploratory framing phase on September 12, marked by the desertion of 10 frames shortly afterward. The exploratory process slowed between September 13 and 17, with only four additional frames being abandoned. By September 17, the category of frames became stable, signaling the end of the trial phase.
The rate of frame desertion.
The result of this exploratory trial was that 17 frames (“Anti-terrorism”“Disastrous Influence”“Threatened Security”“Crisis Alleviation,”“Conspiracy”“Recovery”“Patriotism”“Terrorist Identity”“Charity”“Trauma,”“Increasing Casualties”“Refractory-strict ruler”“Harassed Muslims”“Mass Rescue”“Increasing Refugee”“Defects Reflection”“Essential Man”) became the stabilized frames from September 18.
Decrease of the Number of Newly Emerged Frames
The emergence of new frames followed a clear pattern. In the first phase, 24 frames appeared, among which 10 were consistently used in reports. The remaining frames underwent a process of trial and desertion. In the second phase, seven new frames emerged and were all used consistently. By the third phase (September 18 to October 7), no new frames appeared, indicating the stabilization of frame categories.
This pattern suggests an exploratory stage in the first phase with frequent trial and desertion, a shift toward stabilization in the second phase, and normalization in the third phase. The first phase saw 77.4% of total frames appear, with 32% being abandoned and 45% remaining in use. The second phase witnessed a slowdown in frame trials, with 22.6% of total frames emerging and being consistently used, while 9.68% of frames from the first phase were abandoned. By the third phase, no frames were deserted, and no new ones emerged, marking the completion of the stabilization process. A detailed breakdown of frame emergence and abandonment is provided in Tables C1 and D1.
Stabilization of the Categories of Frames Used Each Day
The stabilization of frame categories reflects the focalization of news coverage. This process is indicated by two key trends: (a) a decrease in the number of frame categories used daily, eventually stabilizing, and (b) an initial increase, followed by stabilization, in the number of reports sharing one frame.
Twenty-four frames were used on the first day. By the second phase, the daily average dropped to 13.6 frames, further decreasing to 11.78 frames in the third phase. This indicates a shift from exploratory framing to more stable categories. Similarly, the number of reports sharing one frame increased from 2.17 in the first phase to 3.89 in the second phase, 2.34 in the third phase. This suggests that while early coverage was dispersed, reporting later concentrated around key frames.
To further elaborate this, we have examined the exact number of frames used each day and the exact number of reports sharing one frame each day. Figure 2 shows the stabilization process of frames.
Stabilization process of framing.
In conclusion, frame category stabilization was marked by a decrease to a steady number of frames used each day, and a trend where the number of reports sharing a frame first rose, then declined to a level higher than the initial one. This pattern can be explained by two factors. First, as some frames were abandoned, more reports clustered around fewer frames, causing an initial rise. Second, large clusters around primary natural frames dissolved, while new clusters around extended frames formed, leading to a decline and eventual stabilization.
Re-Focalization of Frames
Re-focalization of frames refers to the shift in news focus, where initially dominant frames became less central, while previously marginal frames gained prominence. Table 5 shows the re-focalization process.
Evolution of Frames and Key Features.
The framing of the 9/11 attacks evolved in a clear temporal sequence, transitioning from uncertainty and emotional shock to counterterrorism justification and security mobilization. This dynamic progression reflects how journalists made sense of events as new information emerged and political actions were taken. In the immediate aftermath of the attacks, news coverage focused heavily on descriptive and emotional frames. The dominant frames—Disastrous Influence, Crisis Alleviation, Increasing Casualties, Condemnation, and Witness Horror—emphasized uncertainty, physical destruction, rescue operations, and grief. During this stage, journalists relied on primary natural frames to narrate immediate human suffering and emergency responses, reflecting the absence of a clearly defined cause or policy direction. As more information became available and official investigations pointed to terrorism, the narrative gradually shifted. While Disastrous Influence and Crisis Alleviation continued, frames such as Condemnation and Witness Horror declined. At the same time, Conspiracy and Anti-terrorism emerged as key interpretive frames. This stage represents a re-focalization process, during which journalists began to link the attacks to organized terrorism, signaling a cognitive shift from emotional response to problem definition.
A decisive shift occurred on September 18, 2001, when the U.S. Congress passed the Authorization for Use of Military Force (AUMF), formally granting President Bush the legal authority to use force against those responsible for the attacks. Following this legislative mandate, media framing rapidly converged on Anti-terrorism and Threatened Security as dominant frames. These extended frames reframed the event from a national tragedy into an act of war, justifying military retaliation and positioning terrorism as an existential threat. Natural frames such as Crisis Alleviation and Increasing Casualties further declined in prominence. After September 26, extended frames surpassed natural frames in frequency, marking the emergence of a new dominant interpretive paradigm centered on national security and counterterrorism policy.
To assess whether the distribution of major frames changed significantly over time, chi-square tests were conducted across the three phases (see Table 6).
Chi-Square Tests Major Frames Across Three Phases.
Note. p = .0011, indicating a statistically significant difference at the 0.01 level.
The overall test of the four selected frames indicated a significant difference among phases (χ2(6) = 22.19, p = .0011), suggesting notable evolution in the framing landscape. Although the chi-square tests for the two subgroups were not statistically significant at the conventional level—Extended frames (χ2(2) = .79, p = .673) and Natural frames (χ2(2) = 1.30, p = .521)—their descriptive trends reveal opposite trajectories. The Extended frames (Anti-terrorism and Threatened Security) showed a clear upward trend, reflecting an increasing focus on externalized, policy-oriented interpretations of the event. In contrast, the Natural frames (Crisis Alleviation and Increasing Casualties) gradually declined, indicating a shift away from immediate disaster reporting toward legitimizing the anti-terrorist discourse.
The passage of AUMF on September 18 marked a turning point in media discourse. Journalists began framing the attacks not merely as a humanitarian disaster but as a national security crisis requiring military intervention. This demonstrates how institutional actions (such as Congressional authorization) directly influenced journalistic framing priorities, leading to stabilization and convergence around extended frames.
Shift of Frames
As previously discussed, 14 frames were dropped from the first to the final phase. We categorized them into four: frame desertion, transformation, absorption, and convergence.
Frame Desertion
Frame desertion occurred because of the shifting news value. Some frames were closely associated with specific topics, and when those topics lost their news value, the frames associated were deserted. Four frames fit this category: “Crashed Plane,” covering the Pennsylvania crash; “Collapsed Building,” explaining the towers’ structural failure; “Comforting Media,” showing the internet’s role in locating missing persons; and “Timely Assistance,” reporting hospital aid efforts. These deserted frames were used in reports on the first day. Due to the limited availability of news materials, any available content related to the attack was organized into reports. This expedient organization of information was valued. However, as journalists gained greater access to materials, the competition for news value intensified. As a result, these topics and the frames used were discarded.
Frame Transformation
Frame transformation occurs when the same topic remains newsworthy but the interpretive lens shifts due to increased contextual clarity or changing news value. This transformation illustrates how journalists adapt frames in response to increasing certainty and contextual cues, moving from speculative to policy-driven interpretations.
In the early stage, when the consequences of the attack were still unknown, journalists employed the frame “Undeterminable Aftermath” to express uncertainty. As factual information accumulated, this uncertainty no longer carried news value. The frame was therefore replaced by “Disastrous Influence,” which connotatively emphasized concrete economic and social impact. This shift reflects not only the abandonment of a frame but a priming effect in which journalists foregrounded frames aligned with the evolving understanding of the crisis and the emerging policy discourse.
Excerpt (Undeterminable Aftermath):
The aftershocks will be nearly as bad… It appears possible that the attacks will undercut Mr. Bush’s campaign … (Awaiting the Aftershocks, Sept 12)
Excerpt (Disastrous Influence):
The terrorist attacks last week are beginning to deal a blow to state economies … (Attacks Strike Hard at States’ Economies, Sept 19)
Secondly, there was a transformation in defining the perpetrators from “Trained Attackers” to “Terrorist Identity.” In the immediate aftermath, when the motives and affiliations of the attackers were still unclear, journalists used the frame “Trained Attackers” to highlight their technical competence and psychological stability. As more intelligence evidence emerged linking the hijackers to Osama bin Laden, the news discourse shifted to the frame “Terrorist Identity.” This transformation reflects a contextual recalibration of news value: once the attackers’ affiliations gained certainty and political relevance, the frame emphasizing their identity as terrorists—rather than their training—became more salient and primed the public to interpret the incident as an act of war rather than an isolated crime.
Excerpt (Trained Attackers):
… Yet the attacks … were in all likelihood the work of perfectly sane people … placed in training that can last for months … (Attackers Believed to Be Sane, September 12)
Excerpt (Terrorist Identity):
Federal investigators are examining a possible link between the hijackers and operatives for Osama bin Laden … (Bin Laden Operative Is Linked to Suspects, September 18)
Thirdly, there was a transformation from “Witnessed Horror” to “Terrorism Threat.” In the earliest reports, when the scale and intent of the attacks were still uncertain, journalists used the “Witnessed Horror” frame to convey emotional immediacy through vivid imagery and eyewitness testimony. As official confirmation identified the incident as terrorism and political leaders began framing it as an act of war, the news discourse shifted to “Terrorism Threat.” This change reflected a redefinition of context: the same visual tragedy was now reinterpreted as part of a global security narrative. The transformation illustrates how framing and priming interacted—emotional depictions initially evoked empathy, but were later replaced by frames that primed readers to perceive the event as a threat demanding collective defense.
Excerpt (Witnessed Horror):
… Carol Willis … watched as Tower 1 exploded outward …“I still see clouds of black smoke,” she said. “But there are no towers. I can’t believe it.” (Personal Accounts of a Morning Rush That Became the Unthinkable, Sept 12)
Excerpt (Terrorism Threat):
President Bush declared today that the attacks … were “acts of war.” He spent much of the day trying to rally an international coalition for what could become a massive military response. (Bush Labels Aerial Terrorist Attacks Acts of War, Sept 13)
Fourthly, there was a transformation from “Thunderous Impact” to “Terrorism Threat.” In the earliest stage, when the physical force of the attack was the most visible and concrete dimension available to journalists, coverage emphasized its seismic impact, framing the event through measurable shock and destruction. As official discourse defined the incident as terrorism and policy responses emerged, the framing shifted to “Terrorism Threat” to reflect its national security implications. This transformation indicates a shift in news value from the physical manifestation of the attack to its geopolitical meaning, priming audiences to interpret the event as an act of war requiring military response.
Excerpt (Thunderous Impact):
The crash of two planes into the World Trade Center … created shock waves that registered on sensitive instruments meant to monitor earthquakes … (Columbia’s Seismographs Log Quake-Level Impacts, Sept 12)
Excerpt (Terrorism Threat):
The resolution passed by Congress authorizing President Bush to use force in response to last Tuesday’s terror attacks is nearly as wide-ranging as a declaration of war … (Measure Backing Bush’s Use of Force Is as Broad as a Declaration of War, Sept 17)
Fifthly, the frame shifted from “Vulnerable America” to “Terrorism Threat.” In the initial stage, when the scale and intent of the attack were still unclear, journalists emphasized national susceptibility, portraying the United States as exposed to asymmetric threats. As intelligence evidence accumulated and the attacks were officially attributed to bin Laden, the narrative moved toward “Terrorism Threat,” foregrounding military response and counterterrorism strategy. This transformation reflects a shift from a passive depiction of national vulnerability to an active framing that primed audiences to view the event as an organized terrorist campaign requiring strategic action.
Excerpt (Vulnerable America):
… no amount of power can provide protection against an enemy with limited means but ruthless determination … (America the Vulnerable Meets a Ruthless Enemy, Sept 12)
Excerpt (Terrorism Threat):
… waging that fight will require new military thinking and bolder tactics … (U.S. Force vs. Terrorists: From Reactive to Active, Sept 14)
Lastly, there was a transformation from “Hope” to “Increasing Casualties.” In the immediate aftermath, when rescue efforts were still underway and outcomes were uncertain, journalists adopted the Hope frame to emphasize the possibility of survival. As time passed and no additional survivors were found, this frame lost relevance. With rising confirmed deaths, coverage shifted to the Increasing Casualties frame, reflecting a change in news value from potential rescue to human loss. This shift illustrates how framing adjusts in response to evolving situational certainty, priming audiences to transition from optimism to mourning.
Excerpt (Hope):
… two men were pulled alive from beneath 30 feet of debris …“At least there is some hope we can get more people out,” Mayor Giuliani said. (Hopes: Survivors are Found in the Rubble, Sept 12)
Excerpt (Increasing Casualties):
No survivor has been pulled from the debris in eight days … officials have prepared the public for the likelihood that nearly all … have died. (Unwilling to Give Up Yet, Rescuers Find Crevices but No Survivors in Them, Sept 20)
Frame Absorption
Frame absorption refers to frames that were initially dominant but later lost headline status as their information became widely accepted and incorporated into other frames. Rather than disappearing, these frames were embedded as background elements supporting newly dominant interpretive structures. In this study, three frames—Condemnation, Malicious War, and Mass Rescue—were absorbed in this manner.
For example, the Condemnation frame, prominent in the initial reaction, was later subsumed under the Anti-terrorism frame. After September 14, headlines and leads focused on counterterrorism actions, while condemnation appeared only as supporting language within the text. In the excerpt below, the phrase “heinous acts” reflects the residual presence of condemnation, but the overall framing centers on anti-terrorism measures. The Gulf Arab states promised “total support and cooperation”… for the effort to find and punish those responsible … while protesting efforts to link the “heinous acts” that occurred in the United States with Islam. (Six Arab States Offer Support, Sept 24)
Secondly, the Malicious War frame, which initially characterized the attacks as an act of war comparable to Pearl Harbor, was later absorbed into the broader Anti-terrorism frame. After September 13, reports no longer used Malicious War as the main organizing lens. Instead, references to war appeared only as background elements reinforcing the central theme of counterterrorism. This reflects a shift from historical analogy to a security-driven narrative, in which war discourse was subsumed under anti-terror policy framing. One week after two jetliners slammed into the World Trade Center … members of New York’s National Guard … have come to realize that they occupy a unique position in the response to the most deadly terrorist attack in American history. (After a War Starts at Home, the Guard Prepares to Take It Abroad, Sept 18)
Thirdly, the Heroic Death frame, which initially highlighted the sacrifice of firefighters, was subsequently incorporated into the Anti-terrorism frame. After September 12, references to fallen firefighters no longer structured headlines or leads but appeared as background elements within reports centered on counterterrorism action. This reflects a shift from commemorative framing toward security mobilization, where heroism was retained to legitimize anti-terror initiatives. President Bush … approved the plan to drop missiles and bombs … His remarks embraced the 343 New York firefighters believed to have died at the World Trade Center. (President Worked Toward a Decision on Attack-Aid Combination, Oct 7)
Convergence of Frames
Convergence of frames refers to the formation of an implicit link between the dominant frame and marginal frames, wherein the latter serve as the supporting and reasoning basis for the former.
In the first phase, frames were scattered, with minimal interaction among them. The dominant frame, Disastrous Influence, existed independently, with no other frame serving as its reasoning basis. Additionally, there was little interaction among the non-dominant frames.
In the second phase, interactions emerged among the non-dominant frames. Frames such as Increasing Casualties, Crisis Alleviation, Threatened Security, Terrorism Threat, Malicious War, Condemnation, Terrorist Identity, and Patriotism became interlinked with the Anti-terrorism frame, providing justification for anti-terrorism coalitions. These interactions indicated an initial stage of convergence, where certain marginal frames began to serve as the reasoning foundation for an emerging extended frame.
In the third phase, as Anti-terrorism became the dominant frame, there was a direct relationship between the dominant frame and non-dominant ones. Frames such as Disastrous Influence, Increasing Casualties, Threatened Security, reinforced the Anti-terrorism frame by elaborating on aspects like the destruction, casualties, risks to security. These elements strengthened the Anti-terrorism narrative.
The convergence process of frames unfolded in three stages. In the initial stage, frames appeared scattered, with little interaction. In the intermediate stage, some marginal ones became reasoning bases for emerging extended frames. By the final stage, these extended frames gained dominance, and other frames offered supporting arguments. Ultimately, this convergence marked a gradual consolidation of multiple frames around the dominant one.
Discussion and Conclusion
The dynamics of framing were reflected in stabilization, re-focalization, shift, and convergence (see Figure 3).
Dynamic framing model.
Firstly, stabilization occurred as the framing process evolved. In the early phase, multiple exploratory frames were employed, characterized by the trial and desertion. During this stage, the number of frame categories used within a single day was at its highest, while the number of reports sharing the same frame was minimal, indicating a lack of focalization. As reporting progressed, some frames were abandoned. Consequently, the number of frame categories and the daily usage of frames declined, while the number of reports sharing one frame increased. Ultimately, frame categories, the daily application of frames, and the number of reports using the same frame reached a stable state.
Secondly, re-focalization was observed throughout the process. In the initial stage, primary natural frames were dominant. As coverage progressed, these dominant frames gradually lost prominence, while previously marginal frames gained dominance. This shift indicated a process of re-focalization, wherein initially prominent focal points became less emphasized, while new focal points emerged. By the final stage, this re-focalization process was complete and the dominant frames were largely extended frames, though some primary natural frames persisted.
Thirdly, three distinct types of frame shifts occurred. Frame desertion took place as initially relevant information, which held news value at the outset, lost its significance. Frame transformation occurred when multiple frames were initially used to define a topic, but as the certainty of topic definition increased, less precise frames were gradually replaced by those with greater specificity. Frame absorption happened as certain information became widely accepted and less newsworthy, both the information and its associated frames were incorporated into other dominant frames, serving as background context.
Finally, frame convergence emerged as the framing process reached its final stage. Initially, different frames operated independently with minimal interaction. However, as re-focalization progressed, an extended frame from the second phase became the dominant frame, around which other frames converged. These subsidiary frames functioned as a reasoning foundation, reinforcing the dominant frame.
Our study proposes a tentative model for the dynamic focalizing process in news framing. Previous scholarship on news discourse has mainly conceptualized framing as a static outcome. van Dijk’s (1988) research on news discourse identified its hierarchical structure, while studies by Gitlin (1980), Entman (1991, 1993), Gamson (1989, 1992), and Druckman (2011) examined how frames shape the organization of news reports. Building on this body of work, our study advances the understanding of framing by demonstrating that in the coverage of continuously reported breaking news events, framing is not fixed but evolves dynamically over time. As news coverage unfolds, frames are gradually selected, refined, and focalized, leading to increasing coherence and stability in interpretive emphasis.
In sum, the proposed Dynamic Framing Model contributes to framing theory by conceptualizing the evolution of frames as a process rather than an outcome. Beyond its theoretical contribution, this model offers practical implications for understanding media dynamics across different contexts. It can be applied to shorter news cycles, where the focalizing process may occur rapidly and thus remain less observable, and to social media environments, where framing dynamics are intensified by user participation, algorithmic amplification, and accelerated temporal rhythms. Likewise, in highly politicized outlets, the model helps explain how initial plurality in framing may converge toward securitized or ideologically dominant frames, raising normative concerns about interpretive homogenization and its effects on public understanding and policy discourse.
More broadly, this study underscores the importance of recognizing the temporal fluidity of news frames. Static frames seen in conventional reporting may, in fact, represent the endpoint of an underlying dynamic focalizing process. Acknowledging this dynamic dimension enables journalists, analysts, and policymakers to better anticipate how public narratives crystallize and to reflect critically on the democratic implications of frame convergence in the digital age. Future empirical research should further test the model across different media systems and event types to assess its generalizability and explanatory power.
This study has several limitations. While we followed van Gorp’s (2010) guidelines and ensured inter-coder reliability, the identification and naming of frames remain interpretative processes. Coders’ subjectivity, particularly in the final stage of naming frames, cannot be entirely eliminated. More refined instruments are needed to minimize subjectivity. This study relies on reports from The New York Times, a leading national newspaper with strong agenda-setting influence and a distinct editorial orientation. Its emphasis on policy analysis, institutional accountability, and geopolitical implications may have shaped the prominence of frames such as Anti-terrorism and Threatened Security, while more immediate, emotionally driven frames like Witness Horror were de-emphasized as coverage progressed. Although this single-source approach may limit external validity, it provides a representative case of how mainstream media construct dominant narratives in response to national crises. To enhance generalizability, future research could incorporate multiple news outlets with diverse political orientations or media systems, including international and digital platforms, to examine whether the dynamic framing model holds across varying editorial contexts. Another limitation of this study is its focus on front-page coverage, which may privilege dominant frames and overlook peripheral perspectives found in interior or op-ed sections. Future research could test the proposed model using a random sample to evaluate its generalizability and examine whether marginal frames display similar or divergent evolutionary patterns.
Supplemental Material
sj-docx-1-sgo-10.1177_21582440251405175 – Supplemental material for Dynamic and Function of Framing in Journalism: A Model Based on 9/11 News Reports
Supplemental material, sj-docx-1-sgo-10.1177_21582440251405175 for Dynamic and Function of Framing in Journalism: A Model Based on 9/11 News Reports by Jianxin Yang and Pingyan Li in SAGE Open
Footnotes
Appendix A
Appendix B
Usage Frequency of Deserted Frames.
| Phase | Deserted frame | Frequency before desertion |
|---|---|---|
| Phase 1 → Phase 2 (Sept. 12–13) | Witnessed Horror | Used once |
| Hope | Used once | |
| Crashed Plane | Used once | |
| Vulnerable America | Used once | |
| Collapsed Building | Used once | |
| Heroic Death | Used once | |
| Comforting Media | Used once | |
| Timely Assistance | Used once | |
| Trained Attackers | Used twice | |
| Undeterminable Aftermath | Used twice | |
| Phase 2 → Phase 3 (Sept. 13–17) | Malicious War | Used once on Sept. 12, twice on Sept. 13 |
| Thunderous Impact | Used once on Sept. 12, once on Sept. 15 | |
| Condemnation | Used once each on Sept. 12, 13, 14 | |
| Terrorism Threat | Used twice on Sept. 12, four times on Sept. 13, once on Sept. 15, once on Sept. 17 |
Appendix C
Emergence and Normalization of New Frames.
| Phase | Newly emerged frames | Normalization status | Emergence date(s) |
|---|---|---|---|
| Phase 1 (Sept. 12–13) | Anti-terrorism, Disastrous Influence, Threatened Security, Crisis Alleviation, Conspiracy, Patriotism, Trauma, Increasing Casualties, Harassed Muslims, Defects Reflection | Used normally | Sept. 12 |
| Witnessed Horror, Hope, Crashed Plane, Vulnerable America, Collapsed Building, Heroic Death, Comforting Media, Timely Assistance, Trained Attackers, Undeterminable Aftermath, Malicious War, Thunderous Impact, Condemnation, Terrorism Threat | Trial and desertion observed | Sept. 12–13 | |
| Phase 2 (Sept. 13–17) | Mass Rescue, Charity, Refractory-strict ruler, Essential Man | Used normally | Sept. 13 |
| Recovery, Terrorist Identity | Used normally | Sept. 14 | |
| Increasing Refugee | Used normally | Sept. 17 | |
| Phase 3 (Sept. 18–Oct. 7) | No new frames emerged | Categories stabilized | — |
Appendix D
Frame Stabilization Process.
| Phase | Trial and desertion of frames | Frames used normally | Frame stabilization process |
|---|---|---|---|
| Phase 1 | 77.4% of total frames appeared; 32% deserted | 45% of total frames survived | Multiple exploratory frames |
| Phase 2 | 9.68% of Phase 1 frames deserted | 22.6% of total frames appeared and were normally used | Transition from exploratory to formative |
| Phase 3 | No new frames emerged; No frames deserted | All remaining frames were normalized | Fully stabilized categories |
Ethical Considerations
This is an examination of texts and no ethics are involved. Therefore, we have exemption from requiring ethics approval.
Consent to Participate
This article does not contain any studies with human participants performed by any of the authors.
Author Contributions
Jianxin Yang conceived the study, conducted the analysis, analyzed the data, and wrote the manuscript. Pingyan Li collected the data, coded the data and analyzed the data.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by Soft Science Special Project of Gansu Basic Research Plan under Grant No. 25JRZG009. 2024–2026 Gansu Provincial Research Project on the Development of Foreign Language Teachers in Higher Education Institutions under Grant No. Z202409.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Supplemental Material
Supplemental material for this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
