Interactive Media Engagement Predicts Posttraumatic Distress After Community Trauma: A National Repeated Cross-Sectional Study

Article information

Psychiatry Investig. 2026;23(10):1307-1316
Publication date (electronic) : 2026 October 1
doi : https://doi.org/10.30773/pi.2026.0172
1Department of Psychiatry, Kangbuk Samsung Hospital, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea
2Department of Biostatistics, Kangbuk Samsung Hospital, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea
3Workplace Mental Health Institute, Kangbuk Samsung Hospital, Seoul, Republic of Korea
4Media Real Research Korea Co., Ltd., Seoul, Republic of Korea
5Department of Psychiatry, CHA University Ilsan Medical Center, Goyang, Republic of Korea
Correspondence: Sra Jung, MD, PhD Department of Psychiatry, CHA University Ilsan Medical Center, 1205 Jungang-ro, Ilsandong-gu, Goyang 10414, Republic of Korea Tel: +82-31-782-8391, E-mail: srajung890@chamc.co.kr
Correspondence: Sung Joon Cho, MD, PhD Department of Psychiatry, Kangbuk Samsung Hospital, Sungkyunkwan University School of Medicine, 29 Saemunan-ro, Jongro-gu, Seoul 03181, Republic of Korea Tel: +82-2-2001-2214, E-mail: sjcho0812@hanmail.net
Received 2026 May 4; Revised 2026 June 6; Accepted 2026 June 26.

Abstract

Objective

Media engagement is an important pathway for indirect exposure to community trauma, but little is known about how interactive platforms and meaning-making relate to posttraumatic psychological outcomes. This study examined whether different types of media usage, particularly interactive online engagement, predict acute and subacute posttraumatic distress following a communitylevel traumatic event.

Methods

A two-wave nationwide repeated cross-sectional survey was conducted 2 weeks (n=1,703) and 2 months (n=2,220) post-incident, using stratified quota sampling. Depressive, anxiety, and posttraumatic stress (PTS) symptoms were assessed using the Patient Health Questionnaire-9, Generalized Anxiety Disorder 7-item scale, and the Posttraumatic Stress Disorder Checklist for the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition. Analyses included chi-square tests, analysis of covariance with Bonferroni correction, and multivariable regression models.

Results

Interactive engagement correlated with elevated psychological distress at both time points. At 2 months, receiving information through online communities and sharing graphic content predicted greater symptom severity and higher odds of clinically significant depressive, anxiety, and PTS symptoms. Passive exposure to graphics was not associated with elevated distress and was linked to lower symptom prevalence. Personal relevance to victims and systemic attribution were also significant predictors, with media types showing clearer differences at 2 months.

Conclusion

Posttraumatic distress following community trauma appears to be shaped more by patterns of interaction with trauma-related information than by visual exposure. Interactive online engagement may amplify psychological symptoms over time, highlighting the need for public health strategies that address interactive media environments.

INTRODUCTION

The psychological impact of community trauma on residents’ mental well-being has been extensively examined. Often termed collective or mass trauma, it refers to traumatic events that extend beyond directly exposed individuals to the broader community [1,2]. In particular, widespread media coverage has broadened the boundaries of local disasters, exposing individuals to trauma far beyond the immediate vicinity [3]. Consequently, a substantial body of evidence underscores the significant role of indirect exposure via media [1,2,4,5]. In light of these findings, some scholars suggest revisiting the Diagnostic and Statistical Manual of Mental Disorders (DSM) posttraumatic stress disorder (PTSD) criteria to better account for media-mediated indirect exposure [6,7]. Currently, Criterion A of the DSM Fifth Edition Text Revision (DSM-5-TR) only recognizes repeated or extreme exposure to traumatic events only in work-related contexts.

Both the magnitude and the nature of such exposure have been examined in relation to psychological outcomes [8,9]. For instance, Holman et al. [1] demonstrated that extensive daily media consumption predicted greater acute stress than direct exposure following the Boston Marathon bombings, and viewing graphic content depicting blood or injury was linked to increased posttraumatic stress (PTS) symptoms [3]. Traditional media have taken steps to reduce public distress, such as limiting graphic coverage after the 9/11 attacks in the United States [3] and strengthening disaster reporting guidelines to restrict graphic imagery and descriptions following national tragedies in South Korea [10]. However, the evolving media landscape has transformed information exposure from one-way delivery to interactive exchange, allowing active user engagement via social media and online communities [11,12]. Social media platforms, such as Instagram, X (formerly Twitter), and YouTube, are generally characterized by personal network-based, identity-linked interactions. In contrast, online community forums such as DC Inside, FM Korea, and Clien in South Korea tend to facilitate anonymous, topic-centered group discussions on shared interests or social issues [13,14]. These interactive online environments can shape collective interpretations of traumatic events through the repeated circulation and reinforcement of emotionally congruent narratives, a phenomenon often described as the “echo chamber effect.” [15] Recently, research has expanded beyond traditional media to explore the influence of new media—especially social media platforms—in the context of collective trauma [4,16,17]. Yet, despite the increasing prominence of interactive media, few studies have examined how these dynamic modes of media engagement following a community trauma relate to psychological outcomes in the general population.

Prior research suggests that individuals’ interpretations of traumatic events shape their posttraumatic psychological outcomes [7]. Literature indicates that perceived similarity or shared social identity with trauma victims is linked to greater distress. Relihan et al. [18] note, such identification functions as an interpretive framework, transforming a distant tragedy into a self-relevant threat. Additionally, DSM criteria permit a PTSD diagnosis even without any visual exposure, provided the trauma involved a close personal relationship with a victim. This underscores the potential importance of non-visual factors—particularly interpretation and communicative content—in predicting psychological responses to community trauma. A sense of safety is fundamental to trauma recovery, yet media engagement strongly shapes these perceptions. While reliable information reduces uncertainty, repeated exposure to threat-related media can heighten perceived risk and reinforce collective insecurity narratives [4,19,20]. Building on these insights, this study focuses on the premise that media engagement acts as a conduit for both visual exposure and critical non-visual factors, specifically shaping event interpretation.

In 2022, a catastrophic crowd crush during late-night Halloween festivities in Seoul, South Korea, claimed 159 lives. As the incident occurred on a densely crowded street filled mostly with young people, it sparked rapid and unmoderated social media coverage—including uncensored real-time videos and livestreams—alongside intense public discourse and extensive traditional media coverage [21,22]. As the incident occurred without an official event organizer, responsibility was unclear, intensifying public debate. This discourse was particularly intense in South Korea, where institutional compliance is widely believed to guarantee public safety. Consequently, attributing the disaster to “deficiencies in the social safety net” risks damaging this societal belief and destabilizing the sense of safety [22]. In the aftermath, residents reported acute increases in depressive symptoms, anxiety symptoms, and anger, symptoms that, although somewhat attenuated over time, remained elevated for up to 2 months [21]. Building on this context, this study investigates predictors associated with acute stress and, more broadly, the elevated symptom levels and psychiatric disorder prevalence following community-level trauma, using the Seoul Halloween crowd crush as a case study. Crucially, it introduces a novel focus: new media interactivity, specifically distinguishing between social media platforms and online community forums.

This study aimed to identify predictors of increased severity and prevalence of depressive, anxiety, and PTS symptoms. Guided by DSM-5-TR diagnostic criteria, repeated cross-sectional surveys were conducted at two post-event time points—2 weeks and 2 months post-incident—corresponding to the diagnostic windows for acute stress disorder (ASD) and PTSD, respectively. The researchers hypothesized that: 1) individuals obtaining information primarily through interactive online media would report higher psychological distress and prevalence rates; 2) attributing responsibility for the incident to a systemic-level factor would predict higher distress and prevalence; 3) exposure to and sharing of event-related graphics would predict higher distress and prevalence; and 4) the strength of these predictors would diminish over time, from 2 weeks to 2 months post-incident.

METHODS

Participants

This study employed a two-wave repeated cross-sectional survey with nationwide community-based samples. Data were collected at 2 weeks (November 8–11, 2022) and 2 months (November 30–December 1, 2022) post-incident, corresponding to the ASD and PTSD phases, respectively. Push notifications were sent to 41,000 pre-registered panel members aged 19 years and older, who had voluntarily enrolled in the survey platform. While participation was limited to once per wave, respondents could participate in both waves. The survey was self-administered via a mobile application, and participants received TNC coin (cryptocurrency) as compensation. Quota sampling—based on age, gender, and region—was employed to ensure the sample reflected the national population. Once quotas were met, subsequent submissions were excluded based on completion time. Initial response totaled 1,750 at 2 weeks and 2,300 at 2 months. After excluding respondents who were unaware of the incident, the final sample sizes were 1,703 at week 2 and 2,220 at month 2.

The study protocol was approved by the Institutional Review Board of Kangbuk Samsung Hospital (IRB No. KBSMC 2023-01-016). The study was conducted in accordance with the principles of the Declaration of Helsinki and Good Clinical Practice. The requirement for informed consent was waived by the Institutional Review Board, and the analyses were conducted using de-identified data.

Covariables and outcomes

Participant data encompassed various characteristics, including age, gender, area of residence, relevance to the victim, initial news outlets, exposure to graphics, sharing graphics, and attribution. Participants were deemed relevant if they were family, relatives, acquaintances, or direct witnesses of the victim; others were considered not relevant. Initial news outlets referred to the participants’ first source of information, categorized into traditional media (television and newspapers), online communities, social media, online news, or acquaintances/others. Exposure to graphics was defined as having viewed incident-related visual content, such as videos or images. Sharing graphics referred to actively distributing such content through digital platforms (e.g., posting or forwarding via social media or messaging applications). Attribution assessed whether participants assigned responsibility to systemic factors or individuals.

This study focused on depressive, anxiety, and PTS symptoms, as these are the most commonly reported forms of psychological distress following disasters [23]. Depressive symptoms were assessed using the Patient Health Questionnaire-9 (PHQ-9), a validated 9-item measure, that evaluates depressive symptoms over the past 2 weeks on a 4-point Likert scale (0–3), with higher scores indicating greater severity [24]. A cutoff score of ≥10 was used to classify participants as having clinically significant depressive symptoms [25]. The internal consistency of the PHQ-9 is high (Cronbach’s α=0.908). Anxiety symptoms were measured using the Generalized Anxiety Disorder 7-item scale (GAD-7), which assesses anxiety symptoms over the past 2 weeks on a 4-point Likert scale (0–3) [26]. A score of ≥10 was used to identify clinically significant anxiety (range: 0–21) [27]. The GAD-7 demonstrated excellent internal consistency (Cronbach’s α=0.943). PTS was evaluated using the PTSD Checklist for DSM-5 (PCL-5), a 20-item self-report measure of DSM-5 PTSD criteria rated on a 5-point Likert scale (0–4) [28,29]. Higher scores indicated greater PTS symptoms severity. A cutoff score of ≥31 was used to indicate clinically significant PTS symptoms [30]. The PCL-5 showed excellent internal consistency (Cronbach’s α=0.969). Cutoff scores were used to identify clinically significant depressive, anxiety, and PTS symptoms. These classifications were based on validated self-report instruments and should not be interpreted as clinical diagnoses, as structured diagnostic interviews were not conducted.

Statistical analysis

Stratified quota sampling—by age, gender, and region of residence—was used to secure national representativeness. Cell weighting and complex sampling adjustments were used to align each wave’s demographics with the National Resident Registration Demographics of October and November 2022.

This study used frequency and descriptive analyses to summarize participant characteristics and psychological outcomes, chi-square tests to compare covariates between 2 weeks and 2 months post-incident, and multiple linear regression analyses to identify the associative factors of depressive, anxiety, and PTS symptoms at both time points. Given the importance of both symptom severity and diagnostic thresholds in mental health research, multivariable logistic regression analyses were performed to examine the likelihood of clinically significant depressive, anxiety, and PTS symptoms at 2 months, the DSM-5-TR timeframe for PTSD assessment. Adjusted odds ratios (ORs) and 95% confidence intervals (CIs) were calculated to evaluate the effects of the initial news outlets, graphics exposure, and graphics sharing. Analysis of covariance (ANCOVA) was used to compare depressive, anxiety, and PTS symptoms across initial news outlets, followed by Bonferroni-corrected post-hoc pairwise comparisons to identify significant group differences. SPSS 28.0 (IBM Corp.) was employed for statistical analyses.

RESULTS

Participants’ characteristics

Although the two study samples were independent, their demographic distributions—including age, gender, and region of residence—were comparable across both time points. Small but statistically significant differences emerged regarding victim relevance, graphic content exposure, and responsibility attribution. Specifically, participants assessed at 2 months reported higher personal relevance to the victim and greater systemic attribution. Conversely, exposure to graphic images or videos was more common at 2 weeks post-incident. Overall, sample characteristics remained highly consistent between assessments. Traditional media was the most prevalent initial news source (52.0% vs. 52.5%), followed by online news (24.4% vs. 23.2%), social media (13.7% vs. 15.1%), and online communities (6.8% vs. 7.3%). Finally, mean levels of depressive, anxiety, and PTS symptoms showed no significant differences between the time points (Table 1).

Demographic and clinical characteristics of the study participants at 2 weeks and 2 months post-incident

Predictors of symptom severity at 2 weeks and 2 months

Multiple linear regression analyses were conducted to examine the associations between key factors and psychological symptom severity at 2 weeks and 2 months (Table 2). Being closely related to the victim consistently correlated with elevated symptom scores across all outcomes and time points, displaying a stronger effect after week 2 (depression: B=2.52, 95% CI 1.75 to 3.28; anxiety: B=2.29, 95% CI 1.58 to 2.99; PTS: B=9.34, 95% CI 7.04 to 11.63) to month 2 (depression: B=3.47, 95% CI 2.83 to 4.11; anxiety: B=3.12, 95% CI 2.49 to 3.74; PTS: B=12.50, 95% CI 10.55 to 14.46). Sharing graphics was associated with increased symptoms of depression (week 2: B=1.11, 95% CI 0.27 to 1.94; month 2: B=0.84, 95% CI 0.18 to 1.51), anxiety (week 2: B=0.76, 95% CI 0.04 to 1.48; month 2: B=0.93, 95% CI 0.31 to 1.56), and PTS (week 2: B=2.55, 95% CI 0.29 to 4.82; month 2: B=4.27, 95% CI 2.33 to 6.20); whereas exposure to graphics was associated with lower depressive and anxiety symptoms (depression week 2: B=-1.30, 95% CI -2.23 to -0.37; month 2: not significant; anxiety week 2: B=-0.93, 95% CI -1.73 to -0.12; month 2: B=-0.58, 95% CI -1.14 to -0.02) both time points and PTS at week 2 and month 2 (week 2: B=-2.73, 95% CI -5.14 to -0.32; month 2: B=-1.96, 95% CI -3.75 to -0.17). Those who initially received news from online communities showed higher anxiety (B=1.01, 95% CI 0.21 to 1.81) and PTS (B=3.35, 95% CI 1.00 to 5.70) at month 2 and “acquaintance/others” showed lower depression at week 2 (B=-2.05, 95% CI -3.47 to -0.64). Attribution to the government was linked to higher symptoms across outcomes and time points (depression week 2: B=1.29, 95% CI 0.76 to 1.82; month 2: B=1.24, 95% CI 0.78 to 1.71; anxiety week 2: B=0.72, 95% CI 0.25 to 1.19; month 2: B=0.88, 95% CI 0.47 to 1.29; PTS week 2: B=3.35, 95% CI 2.03 to 4.66; month 2: B=3.09, 95% CI 1.91 to 4.28).

Predictors of the level of depressive, anxiety and PTS symptoms at 2 weeks and 2 months post-incident*†

Symptom variation by information source

Adjusted group differences in depressive, anxiety, and PTS symptoms across initial news outlets were examined using ANCOVA (Table 3). At 2 weeks, depressive symptoms (p=0.040) and PTS symptoms (p=0.002) differed significantly by group, while anxiety symptoms did not (p=0.072). Post-hoc analyses indicated that participants who received information through non-media sources (e.g., acquaintances) reported significantly lower symptom levels than those in media-based groups. Specifically, depression scores were significantly higher in the online community (estimated mean [EM]=9.03), online news (EM=8.63), and traditional media (EM=9.00) groups compared to the acquaintances/others group (EM=6.95). Similarly, all media-based groups reported significantly higher PTS symptom levels than the acquaintances/others group (PTS: p=0.002). At 2 months, group differences persisted for PTS (p=0.020), but not for anxiety (p=0.170) or depression (p=0.730). Anxiety levels were significantly lower for traditional media users (EM=5.81) than online community users (EM=6.82). For PTS, symptoms were highest in the online community group (EM=19.98) significantly exceeding online news (EM=16.69) and traditional media (EM=16.31) groups. Additionally, social media users (EM=18.22) reported significantly higher PTS levels than traditional media users.

Comparison of the level of depressive, anxiety and PTS symptoms among initial news outlet groups*

Predictors of depressive, anxiety, and PTS symptoms at 2 months

Adjusted associations with probable depressive, anxiety, and PTS symptoms at 2 months were examined using logistic regression analyses (Supplementary Table 1). Compared to traditional media users, individuals who obtained information through online communities had significantly higher odds of clinically significant depressive symptoms (adjusted odds of depression [aOR]=1.63, 95% CI: 1.12 to 2.38), and anxiety (aOR=1.59, 95% CI: 1.03 to 2.48). Graphic exposure was associated with lower symptom prevalence. Specifically, those exposed to graphics exhibited lower odds of depression (aOR=0.70, 95% CI: 0.53 to 0.94) and anxiety (aOR=0.67, 95% CI: 0.48 to 0.93) than those not exposed. Those who had shared graphics showed substantially higher odds of clinically significant PTS symptoms (aOR=2.07, 95% CI: 1.46 to 2.93) than those who had not.

DISCUSSION

This study investigated how different forms of media exposure and engagement following a community trauma relate to posttraumatic psychological symptoms at two clinically distinct time points—the ASD phase at 2 weeks and the PTSD phase at 2 months. Using repeated nationally representative cross-sectional samples with complex sampling weights, both symptom severity and the prevalence of clinically significant depressive, anxiety, and PTS symptoms were examined. This study’s key strength is that it differentiates between new media types (online communities, social media, and online news) and traditional media. It also separates passive exposure (viewing incident-related graphics) from active engagement (sharing content). By integrating linear, logistic, and group comparison analyses, it facilitates identifying distinct pathways linking media-related behaviors to psychological outcomes, demonstrating that the nature of media interaction—rather than mere visual exposure—drives posttraumatic responses.

At both assessment points (2 weeks and 2 months), mean levels of depressive, anxiety, and PTS symptoms remained relatively stable, indicating that psychological distress may persist beyond the acute posttraumatic period. This pattern aligns with prior population-level disaster studies reporting that posttraumatic symptoms often plateau rather than rapidly decline, even months after exposure, despite heterogeneity in individual trajectories [31]. Among all covariates, personal relevance to the victim consistently showed the strongest association with elevated symptom levels. This association became more pronounced at 2 months post-event, suggesting that proximity-based distress can intensify rather than diminish over time. This pattern aligns with trauma research demonstrating that direct exposure, personal involvement, or perceived proximity to victims are among the most robust predictors of both acute stress and longer-term posttraumatic psychopathology, often following a dose–response relationship in which greater exposure predicts greater distress [32]. Prior work further suggests that these effects may reflect heightened emotional salience, intrusive re-experiencing, and a reduced capacity for cognitive distancing when a traumatic event involves close others [33]. Taken together, these findings indicate that persistent symptomatology following community trauma is not uniformly distributed, but disproportionately concentrated among individuals with personal relevance or proximity to the event. This pattern underscores the need for early identification and targeted support for these high-risk groups, consistent with established disaster mental health frameworks that prioritize directly exposed or bereaved individuals for follow-up and intervention [34,35].

Patterns related to the initial news outlets revealed a differentiated picture over time. Two weeks post-incident, symptom differences were driven primarily by individuals who first learned of the incident through acquaintances. This group showed lower levels of depressive, anxiety, and PTS symptoms compared to media-based groups. However, differences among the specific media types were negligible during this early phase. By 2 months, clear distinctions emerged: individuals who first received information via online communities displayed significantly higher PTS symptoms, as well as a higher prevalence of clinically significant depressive and anxiety symptoms compared to traditional media users. In contrast, traditional media users showed comparatively lower distress levels, suggesting that the informational environment in which traumatic content is initially encountered may exert delayed but clinically meaningful effects on psychological outcomes. Several communication-related mechanisms described in prior studies may help contextualize these findings. Traditional media and online news typically provide fact-based and vetted information delivered in a structured, largely unidirectional format. Such environments support informational coping and controlled exposure, corelating with lower emotional contagion and reduced rumor amplification during crises [36]. Conversely, online communities are participatory platforms driven by usergenerated content, where fragmented, unverified, or emotionally charged information spreads rapidly via comment threads, reposting, and algorithmic visibility. These platforms are linked to heightened rumination, social comparison, and emotional contagion [37,38], and accelerated misinformation diffusion during crises [39,40]. This provides a plausible basis for this study’s finding that online community exposure predicted greater distress at 2 months—when discursive processes and collective meaning-making intensify. Attribution of responsibility emerged as another robust factor shaping posttraumatic responses. As hypothesized, participants who attributed responsibility to a systemic-level actor consistently reported higher symptom levels across all domains at both time points. These findings underscore the role of subjective interpretation and meaningmaking in posttraumatic psychological outcomes, consistent with prior research emphasizing cognitive and attributional processes in trauma-related distress. The lower symptom levels observed among individuals who first heard about the incident through acquaintances may reflect interpersonal filtering or social buffering effects. Interpersonal networks can contextualize information, reduce uncertainty, and provide emotional support, both of which have been shown to mitigate post-disaster psychological distress [41]. However, the small sample size in this category at 2 months warrants cautious interpretation. Finally, the cross-sectional design prevents causal inference, and the possibility that individuals with higher initial distress were more likely to seek out or remain engaged in interactive online spaces—a pattern documented in longitudinal disaster studies—cannot be ruled out [36]. Nevertheless, converging evidence indicates that participatory and emotionally charged media environments may exacerbate distress over time, whereas structured informational environments may not, supporting the plausibility of this interpretation. Taken together, these findings suggest that initial contact with trauma-related information is not neutral: media environments characterized by participatory discourse and unregulated content circulation may contribute to heightened psychological distress, whereas structured or socially supported pathways may confer relative protection.

A notable pattern emerged regarding visual content: passive exposure to graphics did not predict greater distress, whereas active sharing of incident-related images and videos did. This divergence suggests that merely seeing graphic material may not be the critical factor in posttraumatic reactions. Instead, the psychological processes embedded in active engagement may be more consequential. Sharing visual content entails intentional selection, re-experiencing, and re-transmission of trauma-related material, which may heighten cognitive and emotional involvement through mechanisms such as repeated mental rehearsal, identification with victims, moral positioning, or a perceived responsibility for information dissemination [39,42]. These processes align with trauma theory emphasizing that re-exposure, emotional amplification, and the absence of cognitive distancing can exacerbate intrusive symptoms and anxiety over time. Experimental and observational work on digital media similarly indicates that participatory behaviors—such as commenting, reposting, and creating content—are more strongly linked to emotional contagion, rumination, and affective arousal than passive browsing or viewing [37]. This interpretation also helps explain why passive exposure was counterintuitively associated with lower distress in this study’s sample. In contemporary media environments where graphic content is pervasive and difficult to avoid, those who report no exposure may represent individuals deliberately avoiding such content as a coping strategy, potentially reflecting higher baseline distress or anxious avoidance. Prior research has shown that disaster-related anxiety is associated with media seeking, whereas depressive symptoms correlate with media avoidance, supporting the possibility that non-exposure reflects psychological withdrawal rather than resilience [43]. Under this framework, exposure itself is not protective; rather, non-exposure may serve as a behavioral marker of avoidance among more distressed individuals. Repeated exposure to graphic material may also contribute to partial emotional habituation or desensitization in some individuals, although this interpretation remains speculative given the cross-sectional design [44,45]. Taken together, these findings suggest that the psychological impact of graphic material lies less in the content itself and more in the nature of engagement surrounding it. Interactive or participatory media behaviors—whether through sharing, reposting, or discursive engagement—may facilitate processes that sustain or amplify traumatic stress, whereas simple viewing may not. This distinction underscores the need for post-disaster mental health frameworks to consider not only what individuals see, but also how they interact with traumatic content online. While this study’s cross-sectional design precludes causal inference, it remains possible that individuals with higher distress are more inclined to engage in sharing behaviors. However, longitudinal evidence linking participatory media use to sustained distress lends plausibility to these pathways. Future studies incorporating temporal sequencing and qualitative analysis of sharing motives may further clarify how digital interaction shapes posttraumatic adaptation.

This study has several limitations. First, the samples at the two time points were independent, and the same individuals were not followed longitudinally. Because timely data collection after an unexpected community trauma was required, following the same individuals over time was not feasible. We therefore used nationally representative repeated cross-sectional samples to examine population-level differences. Nonetheless, the possibility that observed changes over time reflect sample differences cannot be entirely ruled out. Second, participation was voluntary and limited to mobile-app users, with cryptocurrency-based incentives potentially overrepresenting more digitally engaged individuals, although recruitment was stratified to match national distributions. Third, while validated screening tools were used, they were not diagnostic interviews; thus, symptom levels and prevalence rates may not directly correspond to clinical diagnoses. To minimize error, widely accepted self-report screening measures with strong psychometric properties were used. Fourth, participants were categorized according to the medium through which they first received the news. While this variable was included as a proxy for the primary information source, the initial news outlet may not reflect the medium most frequently used. Fifth, media type classification relied on self-report, and the boundary between social media and online communities can be ambiguous, which may have reduced the precision of media classification. Future studies should establish clearer operational definitions. Lastly, the nature and intensity of graphic exposure were not further specified, limiting the interpretation of the unexpected findings regarding lower distress among those exposed.

In conclusion, this study highlights the significant role of interactive engagement and event interpretation in shaping psychological outcomes following community trauma. While passive exposure to graphic media was not linked to higher distress, active forms of engagement—such as participation in online communities and sharing graphics—were associated with increased symptom severity and a higher prevalence of clinically significant depressive, anxiety, and PTS symptoms. Notably, these associations were more pronounced at 2 months post-incident, underscoring the potential for discursive processes and emotional contagion to amplify distress over time. These findings suggest that in the aftermath of a disaster, psychological impacts are linked not only to the nature of media content, but also to the ways in which individuals interact with and interpret that content. Public health responses should, therefore, address the quality and dynamics of posttrauma information exchange, particularly within interactive online spaces, as part of broader strategies to mitigate long-term mental health consequences. In sum, digital interaction and interpretation may contribute to posttraumatic distress as much as direct visual exposure.

Supplementary Materials

The Supplement is available with this article at https://doi.org/10.30773/pi.2026.0172.

Supplementary Table 1.

Predictors of clinically significant depressive, anxiety, and PTSD symptoms at 2 months post-incident†

pi-2026-0172-Supplementary-Table-1.pdf

Notes

Availability of Data and Material

Data necessary to interpret, replicate, and build on the methods or findings reported in this article are available upon request from the corresponding author. Data are not publicly available due to ethical restrictions that protect patient privacy and consent.

Conflicts of Interest

The authors have no potential conflicts of interest to disclose.

Author Contributions

Conceptualization: Sun Wook Jung, Sra Jung, Sung Joon Cho. Data curation: Kwang-yeol Lee, Nahyun Oh. Formal analysis: Mi Yeon Lee. Investigation: Kwang-yeol Lee, Nahyun Oh. Methodology: Sun Wook Jung, Sra Jung, Sung Joon Cho, Mi Yeon Lee. Supervision: Sra Jung, Sung Joon Cho. Validation: Sun Wook Jung, Sra Jung, Sung Joon Cho. Writing—original draft: Sun Wook Jung. Writing—review & editing: all authors.

Funding Statement

This work was supported by the Lee Si-Hyung Social Psychiatry Research Fund (2023) from the Korean Neuropsychiatric Research Foundation.

Acknowledgments

None

References

1. Holman EA, Garfin DR, Silver RC. Media’s role in broadcasting acute stress following the Boston marathon bombings. Proc Natl Acad Sci U S A 2014;111:93–98.
2. Neria Y, Nandi A, Galea S. Post-traumatic stress disorder following disasters: a systematic review. Psychol Med 2008;38:467–480.
3. Holman EA, Garfin DR, Silver RC. It matters what you see: graphic media images of war and terror may amplify distress. Proc Natl Acad Sci U S A 2024;121:e2318465121.
4. Chao M, Xue D, Liu T, Yang H, Hall BJ. Media use and acute psychological outcomes during COVID-19 outbreak in China. J Anxiety Disord 2020;74:102248.
5. Li K, Li J, Li Y. The effects of social media usage on vicarious traumatization and the mediation role of recommendation systems usage and peer communication in China after the aircraft flight accident. Eur J Psychotraumatol 2024;15:2337509.
6. Oz IT, Cona G. Impact of indirect trauma and disaster media exposure on psychological states and temporal processes: the case of 2023 Turkey earthquakes. Clin Psychol Psychother 2024;31:e70008.
7. Hall BJ, Xiong YX, Yip PSY, Lao CK, Shi W, Sou EKL, et al. The association between disaster exposure and media use on post-traumatic stress disorder following Typhoon Hato in Macao, China. Eur J Psychotraumatol 2019;10:1558709.
8. Otto MW, Henin A, Hirshfeld-Becker DR, Pollack MH, Biederman J, Rosenbaum JF. Posttraumatic stress disorder symptoms following media exposure to tragic events: impact of 9/11 on children at risk for anxiety disorders. J Anxiety Disord 2007;21:888–902.
9. Schlenger WE, Caddell JM, Ebert L, Jordan BK, Rourke KM, Wilson D, et al. Psychological reactions to terrorist attacks: findings from the National Study of Americans’ Reactions to September 11. JAMA 2002;288:581–588.
10. Lee HP, Choi YK, Lee JH, Lee HS. [The psychological consequences of indirect trauma exposure through the news on the Sewol ferry disaster]. Korean Psychol J Cult and Soc Issues 2016;22:411–430. Korean.
11. Neubaum G, Rösner L, Rosenthal-von der Pütten AM, Krämer NC. Psychosocial functions of social media usage in a disaster situation: a multi-methodological approach. Comput Hum Behav 2014;34:28–38.
12. Oksanen A, Celuch M, Oksa R, Savolainen I. Online communities come with real-world consequences for individuals and societies. Commun Psychol 2024;2:71.
13. Schwämmlein E, Wodzicki K. What to tell about me? Self-presentation in online communities. J Comput Mediat Commun 2012;17:387–407.
14. Jaidka K, Zhou A, Lelkes Y, Egelhofer J, Lecheler S. Beyond anonymity: network affordances, under deindividuation, improve social media discussion quality. J Comput Mediat Commun 2022;27:zmab019.
15. Cinelli M, De Francisci Morales G, Galeazzi A, Quattrociocchi W, Starnini M. The echo chamber effect on social media. Proc Natl Acad Sci U S A 2021;118:e2023301118.
16. Goodwin R, Lemola S, Ben-Ezra M. Media use and insomnia after terror attacks in France. J Psychiatr Res 2018;98:47–50.
17. Liu N, Ma Z. Psychiatric reactions among the non-exposed population who viewed disaster-related short videos: evidence from the 2021 Henan floods. J Psychiatr Res 2022;150:21–33.
18. Relihan DP, Jones NM, Holman EA, Silver RC. Shared social identity and media transmission of trauma. Sci Rep 2023;13:11609.
19. Jones NM, Garfin DR, Holman EA, Silver RC. Media use and exposure to graphic content in the week following the Boston marathon bombings. Am J Community Psychol 2016;58:47–59.
20. Rozanov VA, Rutz W. Psychological trauma through mass media. World Soc Psychiatry 2021;3:77–86.
21. Jung SW, Jung S, Lee MY, Oh KS, Shin YC, Shin DW, et al. Psychological distress trends and effect of media exposure among community residents after the Seoul Halloween crowd crush. J Korean Med Sci 2024;39:e233.
22. Choi YJ, Namgung HI. Still in there-citizens’ well-being and PTSD after Seoul Halloween crowd crush in Korea: a cross-sectional study. Sci Rep 2024;14:20537.
23. Fan X, Ning K, Ma TS, Aung Y, Tun HM, Thin Zaw PP, et al. Posttraumatic stress, depression, and anxiety during the 2021 Myanmar conflict: a nationwide population-based survey. Lancet Reg Health Southeast Asia 2024;26:100396.
24. Park SJ, Choi HR, Choi JH, Kim K, Hong JP. [Reliability and validity of the Korean version of the patient health questionnaire-9 (PHQ-9)]. Anxiety Mood 2010;6:119–124. Korean.
25. Manea L, Gilbody S, McMillan D. A diagnostic meta-analysis of the patient health questionnaire-9 (PHQ-9) algorithm scoring method as a screen for depression. Gen Hosp Psychiatry 2015;37:67–75.
26. Seo JG, Park SP. Validation of the generalized anxiety disorder-7 (GAD-7) and GAD-2 in patients with migraine. J Headache Pain 2015;16:97.
27. Spitzer RL, Kroenke K, Williams JB, Löwe B. A brief measure for assessing generalized anxiety disorder: the GAD-7. Arch Intern Med 2006;166:1092–1097.
28. Kim WH, Jung YE, Roh D, Kim D, Chae JH, Park JE. Development of Korean version of PTSD checklist for DSM-5 (K-PCL-5) and the short form (K-PCL-5-S). Psychiatry Investig 2022;19:661–667.
29. Kim JW, Chung HG, Choi JH, So HS, Kang SH, Kim DS, et al. [Psychometric properties of the Korean version of the PTSD checklist-5 in elderly Korean veterans of the Vietnam war]. Anxiety Mood 2017;13:123–131. Korean.
30. Forkus SR, Raudales AM, Rafiuddin HS, Weiss NH, Messman BA, Contractor AA. The posttraumatic stress disorder (PTSD) checklist for DSM-5: a systematic review of existing psychometric evidence. Clin Psychol (New York) 2023;30:110–121.
31. Galea S, Ahern J, Resnick H, Kilpatrick D, Bucuvalas M, Gold J, et al. Psychological sequelae of the September 11 terrorist attacks in New York City. N Engl J Med 2002;346:982–987.
32. Kristensen P, Weisæth L, Heir T. Bereavement and mental health after sudden and violent losses: a review. Psychiatry 2012;75:76–97.
33. McNally RJ. Progress and controversy in the study of posttraumatic stress disorder. Annu Rev Psychol 2003;54:229–252.
34. Norris FH, Friedman MJ, Watson PJ. 60,000 disaster victims speak: part II. Summary and implications of the disaster mental health research. Psychiatry 2002;65:240–260.
35. World Health Organization. Mental health in emergencies: mental and social aspects of health of populations exposed to extreme stressors Geneva: World Health Organization; 2003.
36. Holman EA, Garfin DR, Lubens P, Silver RC. Media exposure to collective trauma, mental health, and functioning: does it matter what you see? Clin Psychol Sci 2020;8:111–124.
37. Feinstein A, Audet B, Waknine E. Witnessing images of extreme violence: a psychological study of journalists in the newsroom. JRSM Open 2014;5:2054270414533323.
38. Kramer AD, Guillory JE, Hancock JT. Experimental evidence of massive-scale emotional contagion through social networks. Proc Natl Acad Sci U S A 2014;111:8788–8790.
39. Vosoughi S, Roy D, Aral S. The spread of true and false news online. Science 2018;359:1146–1151.
40. Oh O, Agrawal M, Rao HR. Community intelligence and social media services: a rumor theoretic analysis of tweets during social crises. MIS Q 2013;37:407–426.
41. Kaniasty K, Norris FH. Longitudinal linkages between perceived social support and posttraumatic stress symptoms: sequential roles of social causation and social selection. J Trauma Stress 2008;21:274–281.
42. Rimé B. Emotion elicits the social sharing of emotion: theory and empirical review. Emot Rev 2009;1:60–85.
43. Houston JB, Spialek ML, First J. Disaster media effects: a systematic review and synthesis based on the differential susceptibility to media effects model. J Commun 2018;68:734–757.
44. Bushman BJ, Anderson CA. Comfortably numb: desensitizing effects of violent media on helping others. Psychol Sci 2009;20:273–277.
45. Fanti KA, Vanman E, Henrich CC, Avraamides MN. Desensitization to media violence over a short period of time. Aggress Behav 2009;35:179–187.

Article information Continued

Table 1.

Demographic and clinical characteristics of the study participants at 2 weeks and 2 months post-incident

Variables 2nd week (N=1,703) 2nd month (N=2,220) Statistics p
Age (yr) 48.41±0.41 48.27±0.36 F=0.126 0.881
 20–29 183 (15.9) 291 (15.8) χ2=0.054 >0.999
 30–39 263 (14.7) 387 (14.7)
 40–49 335 (18.2) 474 (18.5)
 50–59 369 (20.1) 542 (20.1)
 ≥60 553 (31.1) 526 (30.8)
Gender χ2=0.004 0.950
 Male 815 (49.1) 1,077 (49.0)
 Female 888 (50.9) 1,143 (51.0)
Region of residence χ2=0.067 0.970
 Seoul 317 (18.9) 450 (19.2)
 SCA-GI 507 (31.7) 662 (31.5)
 Non-SCA 879 (49.4) 1,108 (49.3)
Relevance to the victim χ2=5.373 0.032
 Not relevant 1,506 (87.7) 1,907 (85.7)
 Relevant/witness 197 (12.3) 313 (14.3)
Initial news outlet χ2=7.833 0.164
 Traditional media 917 (52.0) 1,149 (52.5)
 Online community 109 (6.8) 165 (7.3)
 Social media 219 (13.7) 330 (15.1)
 Online news 411 (24.4) 532 (23.2)
 Acquaintances/others 47 (3.2) 44 (2.0)
Exposure to graphics χ2=20.657 <0.001
 Not exposed 180 (11.0) 337 (16.0)
 Exposed 1,523 (89.0) 1,883 (84.0)
Sharing graphics χ2=1.567 0.236
 Not shared 1,491 (87.7) 1,924 (86.2)
 Shared 212 (12.3) 296 (13.8)
Attribution χ2=8.017 0.030
 Individual level 781 (47.2) 953 (44.0)
 Systemic level 886 (50.6) 1,195 (52.6)
 Both/none/others 36 (2.2) 72 (3.4)
Anxiety symptoms
 GAD-7 score 4.78±0.12 4.85±0.11 F=0.189 0.663
Depressive symptoms
 PHQ-9 score 7.09±0.14 7.13±0.12 F=0.055 0.815
PTS symptoms
 PCL-5 score 13.19±0.35 13.48±0.32 F=0.391 0.532

Values are presented as mean±standard deviation or number (%). SCA, Seoul Capital Area; GI, Gyeonggi/Incheon; GAD-7, Generalized Anxiety Disorder-7; PHQ-9, Patient Health Questionnaire-9; PCL-5, posttraumatic stress disorder Checklist for DSM-5; F, F-statistic; p, pvalue; PTS, posttraumatic stress.

Table 2.

Predictors of the level of depressive, anxiety and PTS symptoms at 2 weeks and 2 months post-incident*†

2nd week
2nd month
B SE 95% CI B SE 95% CI
Depressive symptoms (PHQ-9)
 Age -0.02 0.01 -0.04 to 0.00
 Gender, male -0.64 0.24 -1.10 to 0.18
 Region of residence, Seoul -0.80 0.29 -1.36 to 0.24
 Relevance to the victim, relevant 2.52 0.39 1.75 to 3.28 3.47 0.33 2.83 to 4.11
 Initial news outlets, acquaintances/others -2.05 0.72 -3.47 to -0.64
 Exposure to graphics, exposed -1.30 0.47 -2.23 to -0.37
 Sharing graphics, shared 1.11 0.43 0.27 to 1.94 0.84 0.34 0.18 to 1.51
 Attribution, government 1.29 0.27 0.76 to 1.82 1.24 0.24 0.78 to 1.71
 R2 0.054 0.082
 F 7.817 15.191
 p‡ <0.001 <0.001
Anxiety symptoms (GAD-7)
 Relevance to the victim, relevant 2.29 0.36 1.58 to 2.99 3.12 0.32 2.49 to 3.74
 Initial news outlets, online community 1.01 0.41 0.21 to 1.81
 Initial news outlets, acquaintances/others -1.58 0.59 -2.75 to -0.42
 Exposure to graphics, exposed -0.93 0.41 -1.73 to -0.12 -0.58 0.28 -1.14 to -0.02
 Sharing graphics, shared 0.76 0.37 0.04 to 1.48 0.93 0.32 0.31 to 1.56
 Attribution, government 0.72 0.24 0.25 to 1.19 0.88 0.21 0.47 to 1.29
 R2 0.042 0.089
 F 6.314 15.053
 p‡ <0.001 <0.001
PTS symptoms (PCL-5)
 Age -0.07 0.02 -0.12 to 0.02
 Relevance to the victim, relevant 9.34 1.17 7.04 to 11.63 12.50 1.00 10.55 to 14.46
 Initial news outlets, online communities 3.35 1.20 1.00 to 5.70
 Initial news outlets, acquaintances/others -5.96 1.70 -9.28 to -2.63
 Exposure to graphics, exposed -2.73 1.23 -5.14 to -0.32 -1.96 0.91 -3.75 to -0.17
 Sharing graphics, shared 2.55 1.16 0.29 to 4.82 4.27 0.99 2.33 to 6.20
 Attribution, government 3.35 0.67 2.03 to 4.66 3.09 0.61 1.91 to 4.28
 R2 0.087 0.154
 F 9.119 23.857
 p‡ <0.001 <0.001
*

age, gender, region of residence, relevance to the victim, initial news outlet, exposure to graphics, sharing graphics, and attribution of responsibility were adjusted;

†

gender - female, Initial news outlet - Traditional media, Exposure to graphics - Not exposed, Relevance to the victim - Not relevant, Attribution of responsibility - Individuals involved were set as the reference groups for the model;

‡

each p-value represents the overall significance of the corresponding regression model based on the F-statistic, not the significance of each individual predictor.

PHQ-9, Patient Health Questionnaire-9; GAD-7, Generalized Anxiety Disorder 7-item scale; PCL-5, posttraumatic stress disorder Checklist for DSM-5; SE, standard error; CI, confidence interval; F, F-statistic; p, p-value; PTS, posttraumatic stress.

Table 3.

Comparison of the level of depressive, anxiety and PTS symptoms among initial news outlet groups*

2nd week
2nd month
EM (SEM) 95% CI EM (SEM) 95% CI
Depressive symptoms
 Online communities (A) 9.03 (0.59) 7.88 to 10.18 8.64 (0.49) 7.68 to 9.59
 Social media (B) 8.37 (0.58) 7.24 to 9.50 8.07 (0.41) 7.27 to 8.87
 Online news (C) 8.63 (0.53) 7.60 to 9.67 8.24 (0.35) 7.57 to 8.92
 Traditional media (D) 9.00 (0.53) 7.95 to 10.04 8.16 (0.33) 7.51 to 8.81
 Acquaintances/others (E) 6.95 (0.82) 5.34 to 8.56 8.88 (0.33) 8.23 to 9.53
 p 0.040 0.730
 Post-hoc Bonferroni test A>E, C>E, D>E
Anxiety symptoms
 Online communities (A) 6.47 (0.51) 5.48 to 7.46 6.82 (0.43) 5.98 to 7.66
 Social media (B) 6.27 (0.51) 5.27 to 7.28 5.95 (0.35) 5.26 to 6.64
 Online news (C) 6.24 (0.50) 5.34 to 7.14 6.04 (0.30) 5.45 to 6.63
 Traditional media (D) 6.18 (0.47) 5.27 to 7.10 5.81 (0.28) 5.26 to 6.36
 Acquaintances/others (E) 4.60 (0.68) 3.27 to 5.94 6.29 (0.77) 4.78 to 7.80
 p 0.072 0.170
 Post-hoc Bonferroni test A>E, B>E, C>E, D>E A>D
PTS symptoms
 Online communities (A) 18.78 (1.52) 15.80 to 21.76 19.98 (1.25) 17.52 to 22.44
 Social media (B) 16.06 (1.39) 13.34 to 18.78 18.22 (1.02) 16.22 to 20.23
 Online news (C) 16.14 (1.20) 13.79 to 18.49 16.69 (0.83) 15.07 to 18.30
 Traditional media (D) 16.41 (1.15) 14.16 to 18.66 16.31 (0.74) 14.86 to 17.77
 Acquaintances/others (E) 10.36 (1.15) 6.70 to 14.02 15.66 (2.40) 10.95 to 20.36
 p 0.002 0.020
 Post-hoc Bonferroni test A>E, B>E, C>E, D>E A>C, D; B>D
*

age, gender, region of residence, relevance to the victim, and attribution of responsibility were adjusted. Exposure to graphics and sharing graphics were not calibrated as these variables are related to the characteristics of the news outlet used.

EM, estimated mean; SEM, standard error of the mean; CI, confidence interval; p, p-value; PTS, posttraumatic stress.