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

ABSTRACT

Background: Previous research has indicated that continuous exposure to disaster-related information through social media can lead to vicarious trauma. However, scholars have recognized the need for further in-depth research into the underlying mechanisms influencing this relationship.

Objective: The purpose of this study is to investigate the impact mechanism of social media usage on vicarious traumatization in users and analyze the roles of recommendation systems and peer communication.

Methods: This study was conducted with college students in China, focusing on the context of the MU5735 aircraft flight accident in China in which 123 passengers and 9 crew members died. Data were collected through an online questionnaire. The partial least square structural equation modelling (PLS-SEM) method was used to test the data and model.

Results: This study obtained valid responses from 1317 participants. The study findings revealed a significant positive correlation between social media usage(β = 0.180，P < .001), recommendation systems usage (β = 0.172, P < .001), peer communication (β = 0.303, P < .001), and the development of vicarious traumatization. Recommendation systems usage (specific indirect effect  = 0.063, P < .001) and peer communication (specific indirect effect  = 0.138, P < .001) mediated the relationship between social media use and vicarious trauma. Additionally, the impact of peer communication on vicarious trauma was found to be higher compared to the effects of continuous social media use and recommendation system use.

Conclusion: The study found that the use of social media to obtain information about accidents, the frequent pushing of accident information by recommender systems, and the frequent discussion of accidents among peers during unexpected accidents contribute to vicarious traumatization. The study suggests that users’ reduced retrieval of accident information via social media, as well as reduced peer-to-peer discussions about accidents, and social media platforms’ adjustment of recommender system algorithm rules to reduce accident information pushes, may help reduce the likelihood of users experiencing vicarious traumatization.

HIGHLIGHTS

• Social media usage significantly affected college users to develop vicarious traumatization.

• Recommendation systems usage and peer communication significantly affected the development of vicarious traumatization.

• Recommendation systems usage and peer communication mediated the relationship of social media usage and vicarious traumatization.

Antecedentes: Investigaciones previas han indicado que la exposición continua a información relacionada con desastres a través de las redes sociales puede llevar al desarrollo de trauma vicario. Sin embargo, los académicos reconocen la necesidad de realizar investigaciones de mayor profundidad, sobre los mecanismos subyacentes que influyen en esta relación.

Objetivo: El propósito de este estudio es investigar el mecanismo de impacto en la utilización de las redes sociales, en el desarrollo de la traumatización vicaria en los usuarios, y analizar el rol de los sistemas de recomendación que sugieren contenido a los usuarios, en función de sus preferencias pasadas y la comunicación entre pares dentro de este contexto.

Método: Este estudio fue conducido con una muestra de estudiantes universitarios en China, centrándose en el contexto del accidente en el vuelo MU5735 en China, en el cual murieron 123 pasajeros y 9 miembros de la tripulación. Los datos sobre medidas autoinformadas de uso de redes sociales, uso de sistemas de recomendación, comunicación entre pares y trauma vicario se recopilaron a través de una encuesta en línea. Se utilizó el método de modelado de ecuaciones de mínimos cuadrados parciales en modelos de ruta (PLS-SEM) para probar los datos y el modelo.

Resultados: Este estudio obtuvo un total de respuestas válidas de 1.317 participantes. Los hallazgos del estudio revelaron una correlación positiva significativa entre el uso de redes sociales (β = 0.180, P < .001), el uso de sistemas de recomendación (β = 0.172, P < .001), la comunicación entre pares (β = 0.303, P < .001) y el desarrollo de traumatización vicaria. El uso de sistemas de recomendación (efecto indirecto específico  = 0.063, P < .001) y la comunicación entre pares (efecto indirecto específico  = 0.138, P < .001) desempeñaron un rol mediador de la relación entre el uso de redes sociales y el trauma vicario. Adicionalmente, se encontró que el impacto de la comunicación entre pares en el trauma vicario era mayor en comparación con los efectos del uso continuo de redes sociales y del uso de sistemas de recomendación.

Conclusiones: El estudio encontró que el uso de las redes sociales para obtener información sobre accidentes, la frecuente exposición de la información sobre accidentes por parte de los sistemas de recomendación y las discusiones frecuente sobre accidentes entre pares, durante accidentes inesperados, contribuyen a la traumatización vicaria.

背景：先前的研究指出，通过社交媒体持续接触与灾难相关的信息可能导致替代性创伤。然而，学者们认识到有必要进一步深入研究影响这种关系的潜在机制。

目标：本研究旨在调查社交媒体使用对用户产生替代性创伤的影响机制，并分析推荐系统的使用和同伴交流在其中的作用。

方法：东航MU5735航空器事故造成123名乘客和9名机组人员死亡，在此背景下，本研究以中国大学生为样本，通过在线问卷调查收集了数据。采用偏最小二乘结构方程建模（PLS-SEM）方法对数据和模型进行测试。

结果：本研究获得了来自1317名受访者的有效数据。研究结果显示，社交媒体使用（β = 0.180，P < .001）、推荐系统使用（β = 0.172，P < .001）和同伴交流（β = 0.303，P < .001）与替代性创伤的产生之间存在显著正相关。推荐系统使用（特定间接效应 = 0.063，P < .001）和同伴交流（特定间接效应 = 0.138，P < .001）在社交媒体使用与替代性创伤之间起到了中介作用。此外，同伴交流对替代性创伤的影响较持续使用社交媒体和推荐系统的影响更大。

结论：研究发现，在突发事故期间，用户使用社交媒体获取事故相关信息，推荐系统频繁推送事故信息以及同伴之间频繁讨论事故，会促使用户产生替代性创伤。研究建议，用户减少通过社交媒体检索事故信息，以及降低同伴间关于事故的讨论，

社交媒体平台调整推荐系统算法规则以减少事故信息推送，可能有助于降低用户经历替代性创伤的可能性。

KEYWORDS:

• Social media usage
• recommendation systems usage
• peer communication
• vicarious traumatization
• aircraft flight accident

PALABRAS CLAVE:

• Uso de redes sociales
• uso de sistemas de recomendación
• comunicación entre pares
• traumatización vicaria
• accidente aeronáutico

关键词:

• 社交媒体使用
• 推荐系统使用
• 同伴交流
• 替代性创伤
• 航空器事故

1. Introduction

On March 21, 2022, China Eastern Airlines Flight MU5735, carrying a total of 123 passengers and 9 crew members, crashed in Teng County, Wuzhou, Guangxi, resulting in the death of all individuals on board (Chinese airliner carrying 132 people crashes). Following the aircraft flight accident, various media outlets extensively covered the accident. Previous studies have found that continuous exposure to disaster-related information in the media can potentially lead to vicarious traumatization (Lisa Mccann & Pearlman, 1990; Mahamid & Berte, 2020). Research has indicated that the media coverage of the MU5735 aircraft flight accident has triggered a significant number of individuals experiencing symptoms of vicarious traumatization (Qiu et al., 2022). Public health experts have emphasized the need to address the potential issue of vicarious traumatization among the general public in response to this accident (One black box found at crash site, transferred for decoding).

The concept of vicarious traumatization was introduced by McCann and Pearlman (Lisa Mccann & Pearlman, 1990), referring to the psychological abnormality or negative reaction experienced by mental health therapists due to their long-term exposure to trauma or mentally ill patients during counselling sessions and interviews. Such psychological abnormalities often stem from empathy towards trauma survivors, leading to significant distress and even psychological breakdown among the professionals (Lisa Mccann & Pearlman, 1990; Sinclair & Hamill, 2007). According to Figley, McCann and Pearlman, vicarious trauma is thought to have three conditional requirements: empathic engagement and exposure to graphic and traumatizing material, the therapist being exposed to human cruelty, and reenactment of trauma within the therapy process (Adams et al., 2006; Lisa Mccann & Pearlman, 1990). Subsequently, the concept of vicarious traumatization has been widely used by researchers to describe the psychological abnormalities that individuals may develop as a result of indirect exposure to certain traumatic events (Smith et al., 2014; What is Vicarious Trauma? Office for Victims of Crime).

Previous studies have indicated that prolonged exposure to information about violent terrorist attacks, disasters, and other emergencies transmitted through the media leads to adverse mental health outcomes (Hopwood & Schutte, 2017; Silver et al., 2013), thereby inducing vicarious trauma (Thompson et al., 2019; Turnbull et al., 2020). Early research on the media’s impact on vicarious traumatization focused primarily on mass media. For example, Sullender and Holman et al. have found that prolonged exposure to television images of violent events or terrorist attacks is significantly positively correlated with vicarious traumatization (Holman et al., 2014; Sullender, 2010). Subsequently, the impact of the Internet on vicarious trauma has gradually become a hot research direction. Related studies focus on impacts of prolonged exposure of websites, social media (Li et al., 2020; Xu & Liu, 2021), as well as exposure to audio, video, and images on the Internet on users (Turnbull et al., 2020). Overall, previous research findings have deepened our understanding of the relationship between media exposure (form and type) and vicarious trauma, that is the information about violent terrorism, disasters, and natural calamities disseminated in the media exposes users to graphic and traumatizing material. User exposure to distressing information and graphic images related to events can lead to empathic engagement (Lisa Mccann & Pearlman, 1990). Repeated exposure to the distressing information or images that may result in the development of acute stress symptoms or psychological trauma (Feinstein et al., 2014).

Past research has found that the frequency rather than the duration of exposure triggers emotional distress, ultimately leading to psychological trauma (Feinstein et al., 2014). Contemporary social media platforms commonly integrate recommendation systems as a fundamental core feature (Chang & Hsiao, 2013). Previous studies have pointed out that recommendation systems are an important factor influencing the possibility of user information exposure, and it frequently pushes specific news, images, videos and other pages to the target user based on built-in algorithmic rules (Nechushtai & Lewis, 2019; Yang, 2016). However, there are few studies on the impact of recommendation systems on vicarious trauma of social media users. The lack of research is not conducive to our correct understanding of the impact of current social media usage on users’ vicarious trauma. To fill this research gap, we propose to study the following questions: What is the relationship between social media usage and users’ vicarious traumatization? What role do recommendation systems usage play in this relationship?

Interpersonal communication is one of the important factors leading to vicarious traumatization. Direct or indirect contact with disaster/catastrophe eyewitnesses in interpersonal communication (Kellermann, 2001; Sutton et al., 2022), as well as contact with others’ accounts or sharing of information about the event (Isobel & Thomas, 2022; Newman et al., 2019), the emotions and thoughts transmitted during social interactions (Coviello et al., 2014), elicit empathic engagement, and thereby potentially cause individuals to experience vicarious trauma. After the aircraft flight accident, related information spread on social media attracted the attention of many users and may have prompted them to share and discuss with their peers. Peer communication is also considered an important factor in the spread/induction of negative psychological effects in disaster events (Luo et al., 2021). So, what role does peer communication play between social media usage and users’ vicarious traumatization? There is little research on this issue. To fill this gap, we will conduct research on this issue.

This research has a threefold contribution. First, based on empirical data, this study deepens our understanding of how social media usage affects vicarious traumatization by constructing a model of the mechanism of associations. Second, this study extends the understanding of how social media usage impacts vicarious traumatization by identifying important mediators such as recommendation systems usage and peer communication. Third, it provides practical implications for preventing vicarious traumatization caused by social media usage.

2. Methods

2.1. Measures

2.1.1. Social media usage

Social media is a group of Internet-based applications that build on the ideological and technological foundations of Web 2.0, and that allow the creation and exchange of user generated content (Kaplan & Haenlein, 2010). In this study, we referred to previous research on the social media usage preferences of internet users in mainland China and employed six items to measure social media usage. These items included instant social communication (eg., WeChat, QQ), Weibo, community-based platforms (eg., Tieba, Zhihu, Tianya Club), short video platforms (eg., TikTok, Kuaishou), online video streaming platforms (eg., Youku, iQiyi, BiliBili), and self-media platforms(eg., Toutiao, Baijiahao) (Cheng et al., 2016; Dong et al., 2017). Each item was presented following the statement: ‘Please tell us how often you have used the following social media information source to obtain information about the MU5735 aircraft flight accident every day since you became aware of the accident.’ The constructs of social media information sources were measured using a 5-point Likert scale (1 = Never use, 5 = Always).

2.1.2. Recommendation systems usage

Recommendation systems are designed to suggest contents (information, pictures, videos pages) to users depending on previous likes and dislikes, information engagement and interaction, etc. (Ricci & Rokach, 2015). We utilized a scale adapted from Lee and Lee (Ricci & Rokach, 2015) to assess individuals’ usage and reliance on recommendation systems while browsing online. The scale includes three items such as ‘ I use the recommendation systems when browsing information/videos’. Responses were scored on a 5-point Likert scale (1 = strongly disagree; 5 = strongly agree).

2.1.3. Peer communication

Peer communication refers to obvious peer interactions among public (Qin & Men, 2019). Peer communication relies on peer-generated or shared information. Peer communication is vital in public crises as research has found that people tend to believe information received from peers during crisis events (Lin et al., 2016; Youn & Shin, 2019). Additionally, peer communication during crisis events significantly influences individuals’ emotions, cognition, and behaviour related to the crisis event (Luo et al., 2021; Youn & Shin, 2019). The measurement items for peer communication were adapted from the scale used in the study by Luo et al. (2021). We referred to the way the scale of Luo et al.'s (2021) paper was presented about the disaster, adapted it with the name of this event, and did a pre-survey to ensure that the modified scale met the established criteria for reliability and validity (Hair et al., 2021). This scale comprised of three items such as ‘I often discuss the MU5735 aircraft flight accident with my friends/relatives’. Responses were scored on a 5-point Likert scale (1 = strongly disagree; 5 = strongly agree).

2.1.4. Vicarious traumatization

The measurement of vicarious traumatization triggered by social media usage employed a scale adapted from Vrklevski and Franklin’s study (2008), comprising seven items, such as: ‘I have been exposed to sad news and experiences about the MU5735 aircraft flight accident through the media’ and ‘I find myself thinking about these sad messages about the MU5735 aircraft flight accident’. The scale has previously been utilized in the study conducted by Liu et al. (Liu & Liu, 2020). Responses to the items were scored using a 5-point Likert scale (1 = strongly disagree; 5 = strongly agree).

Participants were also requested to provide their demographic information, including gender, age, education level, and current location.

Based on the research questions and the literature discussed above, we hypothesized the following Figure 1:

H1: Social media usage is positively correlated with vicarious traumatization.

H2: Social media usage is positively correlated with recommendation systems usage.

H3: Recommendation systems usage is positively correlated with vicarious traumatization.

H4: Social media usage is positively correlated with peer communication.

H5: Peer communication is positively correlated with vicarious traumatization.

H6: Recommendation systems usage mediates the relationship between social media usage and vicarious traumatization.

H7: Peer communication mediates the relationship between social media usage and vicarious traumatization.

Figure 1. Research model of the impact of social media usage on vicarious traumatization. SMU: social media usage; RSU: recommended systems usage; PeerC: peer communication; VT: vicarious traumatization.

Hypothesis of a direct relationship between social media usage and vicarious traumatization, and mediating roles of recommendation systems usage and peer communication.

2.2. Recruitment

College students are more susceptible to the effects of internet information (Chao et al., 2020; Pat-Horenczyk et al., 2021; Wineburg & Mcgrew, 2016), and they are also a group that is prone to experiencing vicarious traumatization (Turnbull et al., 2020). Therefore, this study focuses on college students as the survey participants.

First, we invited a translator to translate all the original English questionnaires used in this study into Chinese. The translated version was then reviewed by a risk communication specialist and a psychology expert to ensure accuracy and appropriateness of the expression. Next, we conducted a pilot study with 30 college students from Guangxi University. The reliability of the variables were calculated based on the pilot-test data using IBM SPSS 23. In the reliability test, Cronbach’s α of each variable was higher than 0.7 (Cerny & Kaiser, 1977). Subsequently, we commissioned the questionnaire survey to be conducted on the Wenjuanxing platform (www.wjx.com). Wenjuanxing is an online survey organization with a sample database of over 150 million individuals. Many renowned academic journals acknowledge studies that have commissioned surveys through this organization (Peng et al., 2019; Zhu et al., 2020). Participants were informed about the objectives of the study and were aware that the survey was anonymous. Prior to completing the questionnaire, participants were presented with a screening question: ‘Have you been following the information related to the MU5735 aircraft flight accident?’ Only those who chose ‘yes’ were allowed to proceed with the questionnaire; if they chose ‘no’, the survey would be automatically closed. The survey was conducted from 23 March 2022, to 10 April 2022. After removing incomplete responses and questionnaires completed within 30 s, a total of 1317 valid samples were obtained.

2.3. Statistical analysis

This study employed the Partial Least Squares Structural Equation Modeling (PLS-SEM) method for two main reasons. First, the study incorporates both formative indicators (social media usage) and reflective indicators (recommendation system usage, peer communication, vicarious traumatization). Second, this research is exploratory in nature. Nunnally (Nunnally and Bernstein (1994)) suggests that the minimum sample size for SEM analysis should be at least 10 times the number of construct items. With a total of 1317 valid samples, this study meets the sample size requirement. Data analysis and model validation were conducted using SmartPLS 4.0 software.

3. Results

3.1. Descriptive statistics

Table 1 presents the demographic characteristics of the participants. Males accounted for 46.2% (n = 609), while females accounted for 53.8% (n = 708) of the sample. Regarding education, undergraduate students represented 84.5% (n = 1113) of the sample, while graduate students accounted for 15.5%. In terms of age, the majority of participants (n = 1189, 90.3%) fell within the 18–24 age range. In terms of current location, the highest proportion was from other cities, accounting for 47.5% (n = 626), followed by Guangxi (outside Wuzhou), accounting for 32% (n = 422), and the accident route areas (Kunming, Guangzhou), accounting for 12.4% (n = 163).

Table 1. Participants sociodemographic characteristics (n = 1317).

Items	Content	Count	Percentage	Mean	STDEV
Gender	Male	609	46.2	1.54	0.5
	Female	708	53.8
Education	Undergraduate	1113	84.5	2.08	0.47
	Graduate (Master’s and PhD)	204	15.5
Age	Below 18	12	0.9	2.09	0.35
	18–24	1189	90.3
	25–29	101	7.7
	30 and above	15	1.1
Current location	Wuzhou, Guangxi	106	8	2.99	1.06
	Guangxi (outside Wuzhou)	422	32
	Accident route areas (Kunming, Guangzhou)	163	12.4
	Other cities	626	47.5

3.2. Measurement model

The measurement model was assessed for reliability, convergent validity, and discriminant validity. Reliability was evaluated using Cronbach’s α and composite reliability (CR) values (Hair et al., 2021). As shown in Table 2, the Cronbach’s α values for each construct ranged from 0.891–0.928, exceeding the threshold of 0.6. The CR values ranged from 0.914–0.954, exceeding the threshold of 0.8. Therefore, internal consistency reliability and composite reliability are satisfactory.

Table 2. Reliability and validity analysis.

Constructs	Items	Loading	α	rho_a	CR	AVE
RSU	RSU1	0.931	0.928	0.932	0.954	0.874
	RSU2	0.933
	RSU3	0.940
PeerC	PeerC1	0.930	0.896	0.897	0.935	0.828
	PeerC2	0.891
	PeerC3	0.909
VT	VT1	0.751	0.891	0.891	0.914	0.604
	VT2	0.771
	VT3	0.786
	VT4	0.781
	VT5	0.784
	VT6	0.794
	VT7	0.773

The Average Variance Extracted (AVE) and item loadings were used to assess convergent validity (Hair et al., 2021). As shown in Table 2, the AVE values for each construct ranged from 0.604–0.874, surpassing the threshold of 0.5. The loading values of each indicator ranged from 0.751–0.940, exceeding the threshold of 0.7. These results indicate good convergent validity.

Discriminant validity was evaluated using the Fornell-Larcker criterion (Hair et al., 2021). As displayed in Table 3, the convergent validity of the constructs in this study was good.

Table 3. Assessment of Discriminant Validity (Fornell-Larker Criterion).

Constructs	VT	PeerC	RSU
VT	0.777
PeerC	0.438	0.910
RSU	0.329	0.308	0.935

The assessment of formative indicators involved examining collinearity (VIF), the significance of both outer weights, and item loadings on the given constructs (Hair et al., 2021). The assessment results of the formative indicators are presented in Table 4. The VIF values should ideally range between 0.2 and 5. In this study, all formative indicators exhibited VIF values ranging from 1.133–1.336, indicating no issues with collinearity. When evaluating the significance of formative items, the significance of both outer weights (p-values) is initially checked, followed by the examination of item loadings. If necessary, the p-values of the item loadings are then inspected. If the outer weights are not significant, the item loadings are examined. If an item loading is below 0.5, the p-value of the item loading is checked. Items with higher loadings are retained, while others are removed from further analysis. In this study, the outer weight for the ‘Weibo’ construct is 0.090, and the item loading is 0.474. Although the outer weight is relatively low, the p-value of the item loading is 0.000, indicating statistical significance. Therefore, this item can be retained for further analysis.

Table 4. Measurement model of Formative Items.

Construct	Items	Mean	SD	VIF	Weights	T	Loadings	P
Social media	Instant social communication	0.337	0.049	1.229	0.337	6.926	0.658	.000
	Weibo	0.361	0.046	1.133	0.363	7.867	0.608	.000
	Community-based platforms	0.202	0.050	1.336	0.203	4.051	0.593	.000
	Short video platforms	0.395	0.050	1.205	0.397	7.895	0.684	.000
	Online video streaming platforms	0.088	0.053	1.276	0.090	1.700	0.474	.000
	Self-media platforms	0.214	0.050	1.340	0.216	4.331	0.573	.000

3.3. Structural model

First, we utilized the PLS Algorithm to generate path coefficient values (β) and coefficient of determination (R²). Subsequently, we employed the Bootstrapping method (subsamples = 5000) to examine the stability of the model (Hair et al., 2019). The results are shown in Figure 2. Social media usage (β = 0.180, P < .001) was found to have a positive effect on vicarious traumatization, substantiating H1. Social media usage (β = 0.357, P < .001) was significantly positively correlated with recommendation system usage, substantiating H2. Recommendation systems usage (β = 0.172, P < .001) was significantly positively correlated with vicarious traumatization, substantiating H3. Social media usage (β = 0.460, P < .001) was significantly positively correlated with peer communication, substantiating H4. Peer communication (β = 0.303, P < .001) was positively correlated with vicarious traumatization, substantiating H5. Recommendation systems usage accounted for 12.7% of the variance in social media usage, peer communication accounted for 21.2% of the variance in social media usage, while vicarious traumatization accounted for 25.7% of the variance in social media usage, recommendation systems usage, and peer communication.

Figure 2. Structural equation model of the impact of social media usage on vicarious traumatization. SMU: social media usage; RSU: recommended systems usage; PeerC: peer communication; VT: vicarious traumatization; ***：p < .001.

Assessment results of a direct relationship between social media usage and vicarious traumatization, and mediating roles of recommendation systems usage and peer communication.

This study employed the mediation analysis method proposed by Zhu et al. (2020). The results of the mediation analysis are presented in Table 5. The usage of the recommendation system exhibits a complementary mediating effect in the relationship between social media usage and vicarious traumatization, with a specific indirect effect value of 0.063 (P < .001). Peer communication also plays a complementary mediating role in the relationship between social media usage and vicarious traumatization, with a specific indirect effect value of 0.138 (P < .001).

Table 5. Results of the Hypothesis(H1-H5).

Hypothesis	Relationship	β	T	P	Results
H1	SMU→VT	0.180	5.714	.000	Supported
H2	SMU→RSU	0.357	13.355	.000	Supported
H3	RSU→VT	0.172	5.408	.000	Supported
H4	SMU→PeerC	0.460	19.243	.000	Supported
H5	PeerC→VT	0.303	9.782	.000	Supported

Table 6. Results of the mediation analysis.

Hypothesis	Relationship	Specific indirect effects	T	P	Mediation	Results
H6	SMU→RSU→VT	0.063	4.535	0.000	Complementary mediation	Supported
H7	SMU→PeerC→VT	0.138	7.180	0.000	Complementary mediation	Supported

4. Discussion

4.1. Principal findings

This study conducted an online survey targeting college students after the MU5735 aircraft flight accident. With a sample of 1317 valid responses, the SEM approach was employed to investigate the mechanism through which social media usage affects vicarious traumatization. The study also examined the mediation roles of recommendation systems usage and peer communication in the relationship between social media usage and vicarious traumatization. We obtained several important findings from the current study.

What is the influential relationship between social media usage and users’ vicarious traumatization? What role does recommendation systems usage play in this relationship?

For the first question, the more frequently college students are exposed to related information on social media after the aircraft flight accident, the more likely they are to experience vicarious trauma. This finding is consistent with the results of Hall et al.'s study (2019) and Turnbull et al.'s study (2020).

For the second question, this study finds that recommendation systems usage intensifies the effect of social media usage on college students’ vicarious traumatization. According to the modelling results, the more frequently college students use social media, the more frequently they also use recommendation systems, and the more likely they are to experience vicarious trauma. Recommendation systems mediate the impact of social media usage on vicarious traumatization. As mentioned earlier, recommendation systems have become a crucial function of social media. Previous studies have found that many users keep the recommendation feature enabled on social media platforms (Nechushtai & Lewis, 2019). Therefore, recommendation systems can continuously affect users’ information exposure on social media (Mitova et al., 2022). In reality, recommendation systems frequently push information to users based on current trending topics and their browsing history (Yang, 2016), which may contribute to the occurrence of vicarious traumatization (Feinstein et al., 2014). The MU5735 aircraft flight accident was undoubtedly one of the highly popular and captivating topics at that time. College students would receive, click, browse, and even actively search for information related to the aircraft flight accident on social media. These factors would prompt the recommendation systems to continuously push related information (text, images, videos, etc.) to college students. Consequently, college students would experience vicarious trauma due to the continuous exposure to aircraft flight accident-related information, affected by both their use of social media and recommendation systems.

What role does peer communication play between social media usage and users’ vicarious traumatization?

First, peer communication acted as a mediator in the relationship between social media usage and vicarious traumatization. According to the modelling results, college students who frequently use social media to obtain information related to the aircraft flight accident also engage in more frequent discussions about the emergency with their peers. This finding is consistent with the results of Boulay et al.'s study (2002). The more frequent the college students engage in discussions with peers about the aircraft flight accident, the more likely they are to experience vicarious trauma. As previous research has shown, discussions and communication among peers about crisis events can have negative psychological impacts on individuals (Luo et al., 2021), which in turn tends to trigger vicarious trauma (Isobel & Thomas, 2022).

Second, peer communication usage burdens the impact of social media usage on vicarious traumatization. College students may obtain various forms (text, images, videos, etc.) and contents of information related to the aircraft flight accident through social media. This information can trigger feelings of uncertainty and fear about the aircraft flight accident, prompting them to engage in discussions with peers to reduce the sense of uncertainty (Duong et al., 2023; Seo, 2021) or alleviate fears (Oh et al., 2013). However, frequent communication among peers can have both positive and negative consequences. On the one hand, it exposes students to a wide range of information in various forms and contents. On the other hand, negative emotions can easily spread through interpersonal communication about disasters, leading to psychological distress (Luo et al., 2021; Spialek et al., 2019), which in turn can contribute to the development of vicarious trauma.

Our study has obtained an additional finding. The impact of peer communication on vicarious traumatization is greater than the impact of social media usage and recommendation systems usage. At the same time, the mediating effect of peer communication is higher than the mediating effect of recommendation system usage (see Table 5). The reason for this could be that while both social media usage and recommendation systems usage mainly result in vicarious traumatization through continuous and frequent exposure to aircraft flight accident related information, peer communication not only involves continuous exposure to such information but also directly transmits negative emotions. People tend to believe information received from peers during crisis events (Lin et al., 2016; Youn & Shin, 2019), thus promotes empathic engagement between individuals, making it more likely to trigger vicarious trauma among college students (Goff & Smith, 2005).

4.2. Theoretical contributions

This study contributes to the existing literature in two main aspects:

1. The current research has established a link between frequent use of social media during disasters and the development of vicarious traumatization, but there is still a lack of understanding regarding the underlying mechanisms. This study addresses this gap by constructing a dual-path model to examine the impact mechanism of social media usage on vicarious traumatization, and further analyzes the roles of recommendation systems usage and peer communication within this context. In doing so, we provide a complementary contribution to the field of vicarious traumatization research.

2. This study focuses on the impact mechanism of social media usage on vicarious traumatization, specifically in the context of the MU5735 aircraft flight accident, with college students as the sample. It provides a novel research case for investigating factors influencing vicarious traumatization and contributes to the existing research on social media.

4.3. Practical implications

This study also has practical implications for social media users, parents, teachers, social media operators and governments.

(1) The study findings show a significant correlation between social media usage and the development of vicarious traumatization in users. This finding suggests that college students should reduce use social media to retrieve and browse information during times of disasters, in order to prevent the development of vicarious trauma due to frequent exposure to disaster-related information on social media. Family members, friends, and teachers of college students can also provide appropriate reminders about reducing using social media during disaster events.

(2) Peer communication mediates the impact of social media usage on vicarious traumatization. This finding indicates that although people have a habit or need to communicate and share information with peers after a disaster, it is important to reduce the frequency to avoid negative consequences caused by excessive communication.

(3) Recommendation systems usage mediates the impact of social media usage on vicarious traumatization. This finding provides insights into the potential role of recommendation algorithms as an intervention tool to prevent the development of vicarious traumatization. Social media operators and internet regulatory authorities can optimize recommendation algorithms to prevent unrestricted pushing of disaster-related information to users, thereby mitigating the risk of vicarious trauma following disaster events.

5. Limitations and future research directions

This study has several limitations.

(1) The survey data were collected in China, and our study focused on the context of the MU5735 aircraft flight accident, which restricts the generalizability of the results to other countries or contexts.

(2) The data collection method employed in this study was an online questionnaire survey (convenience sampling) targeting college students, which also limits the generalizability of our findings. Future studies could aim to collect data from diverse age groups, educational backgrounds, and occupations.

(3) The data in this study were obtained through self-reported measures, which may introduce subjective biases. Although this method is feasible and widely used due to its convenience and low cost, future research could consider utilizing multiple data sources (e.g. combining self-report data with objective data) to enhance the validity of the results.

(4) This study is a cross-sectional study, which does not allow for the observation of changes in participants’ states over time and makes it challenging to establish causal relationships from social media usage, recommendation systems usage, and peer communication to vicarious traumatization. Future research could consider conducting longitudinal or experimental designs to further explore the causal relationships between these constructs. This would allow for the tracking of changes over time and the testing of causal hypotheses that our study may have generated.

(5) This study measured peer communication based on an existing scale, only assessing whether users frequently communicate with peers, without quantifying the specific frequency. Participants may have varying interpretations of what ‘frequently’ means, which could potentially affect the precision of the results. It is recommended for future research to set more precise measurement items to obtain more accurate data.

6. Conclusions

This study focused on the context of the MU5735 aircraft flight accident and investigated the impact mechanism of social media usage on vicarious traumatization using a sample of 1317 college students who were social media users. The study findings revealed that social media usage, recommendation systems usage, and peer communication are associated with college students’ vicarious traumatization. Meanwhile, recommendation systems usage and peer communication played mediating roles in the relationship between social media usage and vicarious traumatization. The findings of this study contribute to the existing literature on factors influencing vicarious traumatization and social media research. Furthermore, the study provides practical implications for preventing vicarious traumatization resulting from social media usage.

Ethics approval

This study was approved by the Medical Ethics Committee of Guangxi University (GXU-2022-247).

Acknowledgements

Authors’ Contributions: KL: Conceptualization, Methodology, Project administration, Resources, Supervision, Validation, Writing—original draft, Writing—review and editing. JL: Formal analysis, Visualization, Writing—original draft, Writing—review and editing. YL: Data curation, Investigation. All authors have read and agreed to the published version of the manuscript.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability

The data presented in this study can be provided upon request to the corresponding author. For ethical reasons, these data cannot be made public.

Additional information

Funding

This work was supported by the Science and Technology Department of Guangxi Zhuang Autonomous Region under Grant number GuikeAD20159083.