Save the Day: How the Dissemination of Tourism Crises Can Reinvigorate a Tourism Destination Image after the Seoul Halloween Crowd Crush

Abstract

For tourism crises, social media present a double-edged sword: while disseminating the damage caused by tourism crises, it also has the potential to enhance the destination’s visibility and serve as a promotional tool. So, we cannot entirely negate the dissemination of tourism crises but rather proactively utilize its positive aspects to promote the sustainable development of the tourism destination image (TDI). Crisis events caused by management mistakes can be particularly damaging to people’s trust in destination management organizations (DMOs), and a crowd crush is a common and highly damaging type of tourism crisis caused by public management mistakes. Thus, the purpose of this study is to summarize the factors that may positively influence TDI in the dissemination of tourism crises such as the Seoul Halloween crowd crush. This study takes the Seoul Halloween crowd crush as an example and constructs a theoretical model based on information ecology theory. The relationships of variables in the model were analyzed through structural equation modeling. The results show that information transparency, subject authority, and social media interactivity positively influence an organic cognitive image. Social connection and social media interactivity positively influence this organic affective image. Finally, this study provides targeted recommendations for DMOs, which is important for the restoration of TDI after crises.

1. Introduction

Tourism crises are unforeseen events that influence tourists’ confidence in a destination and disrupt normal destination operations [1]. The dissemination of tourism crises on social media can influence a tourism destination image (TDI) [2]. Tourism crises may lead people to question the capabilities of destination management organizations (DMOs), thereby resulting in a boycott of tourism activities. Social media will amplify the dissemination of such emotions and opinions, exacerbating people’s negative perceptions [3]. Thus, the dissemination of tourism crises on social media can negatively influence a TDI [4]. There have been many studies that have confirmed the existence of this negative impact [5,6]. But social media are a double-edged sword [7]; it can also enhance the visibility of a destination when it spreads the damage caused by a crisis event [8], and measures to handle the crisis properly can instead further enhance tourists’ trust in the destination [1,9]. TDI is resilient, and under the impact of tourism crisis dissemination, it also possesses ample adjustability [10]. It has been proven that there is a tipping point in the negative emotions caused by the dissemination of tourism crises. Rather than worsening continuously, the public’s negative emotions will gradually reverse [11,12]. Thus, we should recognize that the dissemination of tourism crises on social media does not only have negative effects. If the factors that can play a positive role are reasonably utilized, it can turn a tourism crisis into a development opportunity, provide a positive reference for the handling of tourism crises, reduce the harm of tourism crises from the aspect of information dissemination [13], and guide the public opinion after the occurrence of tourism crises in a targeted manner, so that TDI can be repaired and enhanced. The types of tourism crises are diverse and complex, with the most widely recognized classification dividing them into two categories: unpredictable crises such as natural disasters and terrorist attacks and crises caused by human management mistakes [1,14]. Because tourism crises caused by management mistakes are more easily avoidable, people are more likely to attribute responsibility to DMOs when such crises occur [15,16]. Consequently, after dissemination, people are more prone to developing negative emotions towards the tourism destination. Thus, we chose to take tourism crises caused by management mistakes as the research object and chose the Seoul Halloween crowd crush as the case of this study, because crowd crush is a common and highly damaging tourism crisis caused by ineffective public management [17]. To our knowledge, there is still no research that systematically illustrates the positive impact of tourism crisis dissemination on TDI, so the main concern of this study is the following:

What factors exist in the dissemination of tourism crises, such as the Seoul Halloween crowd crush, on social media that can positively influence TDI?

The Seoul Halloween crowd crush is a severe stampede that occurred in Itaewon of South Korea. The event resulted in significant loss of life and property, leading to the punishment of responsible parties for negligence resulting in death [18]. This event seriously disrupted the operation of the local tourism industry and compromised the safety of tourists’ activities, representing a typical tourism crisis resulting from management mistakes [19]. Meanwhile, the event has sparked heated discussions among people and has been widely disseminated on social media, which could help us better understand the impact of its dissemination on people’s perception and provide convenient conditions for a questionnaire survey [20]. Thus, we consider that this event can provide a deeper understanding of the impact of tourism crisis dissemination on TDI and provide effective evidence for tourism destinations to manage the dissemination of tourism crises.

To understand the dissemination of tourism crises on social media in a more systematic way, this study introduces information ecology theory as an overall analytical framework. Information ecology refers to a system structure that is formed around information flow and information mapping of the information subject, information, and information environment together [21]. The dissemination of tourism crises on social media can be understood as the following: the events are transcribed into online information elements, which then enter the information ecosystem of social media, thus influencing components of the information ecosystem such as people [22]. Thus, by introducing the theory of information ecology, the dissemination scenario of tourism crises on social media can be structured, allowing for a more comprehensive consideration of various factors in information dissemination and a better understanding of its dissemination patterns. Based on the systematic perspective provided by information ecology theory, we categorized the factors that may positively influence TDI in the dissemination of tourism crises into three dimensions: information, information subject, and information environment.

In terms of how TDI is shaped, the impact of tourism crisis dissemination on TDI primarily centers around an organic image [23]. An organic image refers to the destination impression formed by individuals through education or the mass culture, public media, literature, and other non-commercial marketing information sources. The process of people being influenced by online public opinion is not a process of proactively collecting destination-related information after generating tourism motivation, but a process of passively being influenced, which is more in line with the concept of an organic image [24].

Thus, in this study, we extracted the factors that positively influence organic image through the dissemination of tourism crises from three dimensions of information ecology: information, information subject, and information environment. Relevant hypotheses were then proposed and verified through structural equation modeling (SEM). By doing so, tourism destinations can proactively engage in the dissemination of tourism crises and capitalize on the positive aspects of dissemination. This enables them to manage online public opinion of tourism crises in a more targeted manner and promote the sustainable development of TDI.

2. Literature Review

2.1. Information Ecology Theory

Information ecology was used at its inception to describe the impact of information technology on the flow of information within organizations [25,26]. Later, Davenport applied information ecology to the management of complex information and documents, suggesting that it is more practical to manage information from an ecological perspective [21]. This view made subsequent scholars pay more attention to the logic of information flow in information ecology [27,28]. On this basis, information ecology is defined by Nardi and O’Day as a system of people, work, values, and technology in a given environment in which people, information, and the environment interact with each other to form a dynamic equilibrium [28,29]. Thus, many scholars divide information ecology into three constituent elements: information, information subject, and information environment [30,31,32,33]. Information refers to a wide range of transferable data, knowledge, opinions, viewpoints, etc.; the information subject refers to the participants or entities involved in the process of information production, transfer, acquisition, etc.; and the information environment refers to the social, cultural, and technological contexts that house or shape information and information subjects.

Information ecology theory has now been widely applied to the analysis of information dissemination on social media and plays an important role in exploring the mechanism of information dissemination and the dynamic evolutionary law of public opinion [34]. For example, Johri et al. analyzed engineering students’ use of digital media and devices (or digital technologies) based on information ecology as a framework [35]. Arriagada et al. drew on information ecology to examine how the evolution of social media platforms shapes the activities and interpretive processes of content creators [22]. Luo analyzed from an information ecology perspective how the role of social bots as information subjects influences information characteristics represented by topic inclination and emotion spreading as well as environmental characteristics represented by information diffusion and network influence [36]. Information ecology played a guiding role in all these studies, suggesting that a good research framework for analyzing complex and diverse social media information dissemination can be provided through information ecology.

Building upon the preceding discussion, the dissemination of information and the interaction between users on social media can form an organic and unified information ecosystem of online public opinion. Thus, this study builds a research framework based on information ecology theory. Specifically, in this study, we constructed a research model that includes three dimensions, information, information subject, and information environment, from which the factors that positively influence the organic image of tourism destinations by the dissemination of tourism crises on social media were extracted. The dimension of information refers to the information content closely associated with the dissemination of tourism crises. These contents cover all facets of a tourism crisis event, encompassing a detailed description of the event, the behaviors and responses of the involved parties, opinions, and comments from the broader community, among other aspects. The dimension of an information subject refers to the entities involved in disseminating tourism crises. This includes individuals who share information and convey their attitudes and opinions via information technology within a specific temporal and spatial context, encompassing ordinary users, governmental bodies, we-media, and others. The dimension of an information environment refers to the attributes of the online platforms offered by social media during the dissemination of tourism crises. By integrating three dimensions, information, the information subject, and the information environment, it became feasible to structure the dissemination of tourism crises on social media, thereby facilitating the derivation of more systematic conclusions.

2.2. Organic Image of Tourism Destination

A tourism destination image (TDI) is the psychological impression and perception of a destination by tourists [37]. Lai and Li expanded the concept of TDI by defining TDI as the psychological experience of a destination by tourists, which includes sensations, perceptions, mental representations, awarenesses, memories, and attitudes [38]. In terms of how TDI is formed, Fakeye and Crompton classified TDI as an organic image, induced image, and a complex image, as shown in Figure 1. Organic image refers to the impression of a destination formed by individuals through education or non-commercial marketing information sources such as the mass culture, public media, and literature, and is passive in nature. Induced image refers to the image generated by conscious advertising, promotion, and publicity push influences of the destination, and is active in nature. Complex image refers to a more integrated image formed by tourists after traveling on the ground at the destination, through their own experiences, combined with their previous knowledge [24,39]. Given these three conceptions of TDI, the dissemination of tourism crises on social media primarily impacts tourists’ perception of TDI passively, mainly affecting the organic image. Thus, in this study, TDI specifically refers to an organic image.

In addition to categorizing TDI in terms of its formation process, there is also a horizontal structure of TDI based on function, as shown in Figure 2. This structure was proposed by Baloglu and McCleary, who categorized the components of TDI into a cognitive image, an affective image, and overall image [40]. Cognitive image refers to people’s rational understanding of the destination based on their previous knowledge, information, and experience, such as knowledge of the destination’s geographic location, cultural heritage, natural landscape, and history. Affective image refers to people’s perceptual feelings and emotional feelings about the destination, which reflects tourists’ emotional expectations of the destination, such as pleasant, relaxing, and excited. Overall image is the comprehensive evaluation of the destination by tourists after considering both cognitive and affective factors. Cognitive image positively influences affective image. Both cognitive image and affective image together positively influence the overall image [41].

In summary, the research object of this study is the organic image of tourism destinations, and the organic image can be deconstructed into an organic cognitive image, organic affective image, and organic overall image. Organic cognitive image refers to people’s rational understanding of the destination based on their previous knowledge, information, and experience before reaching the destination. Organic affective image refers to people’s emotional expectations of the destination before generating tourism motivation. Organic overall image refers to people’s overall evaluation of a destination based on a combination of cognitive and affective, non-initiative evaluations of the destination.

2.3. Tourism Crisis and Social Media

A tourism crisis is broadly defined as any event that may threaten the normal operations and behaviors of tourism-related businesses [42]. Scholars later added a new defining condition to the concept of a tourism crisis, namely the reaction of tourists. So, a tourism crisis is explained as an unanticipated event that influences tourists’ confidence in a destination and disrupts the continuation of normal operations of the destination [43,44]. Tourism crises damage the overall reputation of a destination in terms of safety, attractiveness, and comfort, negatively affecting tourists’ perceptions of the destination [1]. Consequently, this leads to an economic downturn in the local tourism industry and interrupts the continuity of business operations due to a reduction in the number of tourists and their expenditures [45].

The widespread adoption of information and communication technology has greatly changed the dissemination patterns of tourism crises and how people engage in spreading information about tourism crises [46]. Before the widespread adoption of the internet and social media, the dissemination of tourism crises mainly relied on traditional media such as television, newspapers, and radio. Public participation in spreading information about tourism crises was relatively limited, typically involving passive reception of information. Nowadays, social media have become the primary platforms for people to access and share information. After a tourism crisis occurs, people can obtain relevant information in real time through social media and exchange views and opinions with others. This two-way, diversified mode of dissemination has expanded the scope and impact of tourism crisis dissemination. People play a more significant role in the dissemination process and are also influenced by the dissemination of tourism crises to a greater extent [47]. In this context, the dissemination of tourism crises on social media influences people’s reactions, attitudes, and behaviors, thereby affecting their perception of a TDI [48,49].

Several scholars have focused on the impact of tourism crisis dissemination on social media on TDI, and most of these studies have demonstrated that dissemination of tourism crises undermines perceptions of the destination and willingness to travel. For example, Xie et al. constructed a model of crisis frames for public online dissemination and found that a crisis frame has a dynamic impact on negative travel intentions [50]. Using a political protest that took place in Hong Kong as an example, Luo and Zhai found that the topics related to it on social media gradually shift from political events to boycotts of tourism activities [51]. However, beyond the damage caused by crisis events, the positive aspects disseminated on social media, including the destination’s crisis management measures, efforts and determination to rebuild its image, and social assistance, help to restore trust in the destination and thus repair its TDI [52]. So, instead of rejecting the dissemination of tourism crises on social media, we need to look objectively at the factors in it that may have a positive impact on TDI. For example, Li et al. used the North Korea nuclear test event spanning from 2016 to 2017 as a case study. They discovered that despite the widespread dissemination of the nuclear crisis, which largely resulted in negative evaluations of North Korea among Chinese tourists, it did not entirely deter their desire to travel. In fact, it added an element of mystery to North Korea in the eyes of tourists, consequently fueling their interest in visiting the country [53]. Zhai et al. found that there is a tipping point for negative emotions triggered by tourism crises during social media dissemination, rather than getting worse all the time, and that negative public emotions can be mitigated by positive and detailed responses prior to, or at the time of, an outbreak [11]. Hence, it is imperative to examine the positive aspects of tourism crisis dissemination on TDI. This exploration can enable destinations to engage more proactively in disseminating tourism crises, mitigate the adverse effects of such crises, or even turn them around. Furthermore, leveraging the positive elements of tourism crises can facilitate the creation of more conducive conditions for the sustainable development of TDI.

3. Research Model and Hypothesis Development

3.1. Information Dimension

3.1.1. Information Transparency and Organic Cognitive Image

Information transparency refers to the extent of openness and clarity demonstrated by DMOs when disseminating information on social media following the outbreak of a tourism crisis. It has been demonstrated that information transparency can significantly reduce the perception of risk and uncertainty toward a tourist destination and increase its perceived value and willingness to act [54].

Information transparency is an attribute that people pay a lot of attention to when facing a public crisis. After the occurrence of a tourism crisis, if tourism destinations can disclose relevant information to the public in a timely, truthful, and comprehensive manner, such information will help to reduce the public’s misunderstandings and panic, thus alleviating the negative impacts of the crisis on TDI [55]. In addition, information transparency can also help the public understand the response measures and improvement plans of the destination, thus enhancing their confidence and satisfaction with the destination. Thus, a first hypothesis is proposed as the following:

Hypothesis 1a (H1a).

Information transparency positively influences organic cognitive image.

3.1.2. Emotional Sympathy and Organic Affective Image

When tourism crises are disseminated on social media, the information often contains emotionally charged content, particularly eliciting emotional sympathy. This emotional sympathy encompasses empathy for the victims, as well as a shared sense of their pain and plight [56]. Emotional sympathy has become a key strategy employed by numerous crisis management organizations [57]. This implies that by establishing an emotional bond with the public, destinations can better understand and address their emotions, mitigating potential resistance from tourists towards the destination following a crisis. Consequently, individuals may become more receptive to the destination’s messaging and initiatives, thereby fostering greater emotional support for the destination. Thus, another hypothesis is proposed as the following:

Hypothesis 1b (H1b).

Emotional sympathy positively influences organic affective image.

3.2. Information Subject Dimension

3.2.1. Subject Authority and Organic Cognitive Image

Subject authority refers to the degree to which the individual or entity publishing a message is recognized and trusted by others in tourism crisis dissemination, representing their leverage on social media [58]. Thus, a subject authority can increase people’s perception of information source credibility, thereby increasing their confidence in subsequent processes such as tourism crisis management and control. It has also been demonstrated that the information source credibility can increase people’s evaluation of TDI [59]. Thus, a second hypothesis is proposed as the following:

Hypothesis 2a (H2a).

Subject authority positively influences organic cognitive image.

3.2.2. Social Connection and Organic Affective Image

Social connection refers to the social relationships that people build through social media in tourism crisis dissemination [60]. Social connection reflects the extent to which people interact and share in tourism crisis dissemination. One of the key insights of social capital theory is that social capital resources are embedded in the social networks of interconnected individuals, groups, or peoples and can be accessed through networks of social relationships [61]. The connections that people make in the dissemination of tourism crises can also be seen by individuals as a form of social capital, and the value placed on social capital can lead to an emotional attachment to the dissemination of tourism crises on social media. This process may lead people to attribute this source of capital to the destination and thus be willing to make more positive comments about the destination. From another perspective, according to social bond theory [62], stronger social bonds can lead individuals to adhere more closely to social norms and reduce aggressive behaviors. Consequently, this may result in more favorable evaluations of the tourism destination. Thus, an additional hypothesis is proposed as the following:

Hypothesis 2b (H2b).

Social connection positively influences organic affective image.

3.3. Information Environment Dimension

3.3.1. Social Media Stability and Organic Cognitive Image

Social media stability refers to the ability of social media to maintain the stability of information dissemination in a tourism crisis situation [63]. Social media stability is directly related to whether credible and reliable information can appear on social media [64]. If social media are unstable, content such as rumors and false information may cause problems such as information distortion and information leakage on social media, leading to an imbalance of the information ecology on social media [65]. For the dissemination of tourism crises on social media, social media stability ensures that people can have continuous access to information about the progress of the handling of tourism crises and reduces their uncertainty about tourism crises. Meanwhile, social media stability can also promote the emotional stability of users, reduce their emotional impulse when people are exposed to information related to tourism crises, and make people more rationally view the impact caused by tourism crises. Thus, a third hypothesis is proposed as the following:

Hypothesis 3a (H3a).

Social media stability positively influences organic cognitive image.

3.3.2. Social Media Interactivity and Organic Image

Social media interactivity refers to the degree to which users interact and communicate with each other and with the content available on the platforms. This interactivity fosters an environment that encourages active participation in information dissemination, facilitating behaviors such as commenting and sharing among users within the information ecosystem. It has been shown that the interactive atmosphere of social media positively influences users’ information-sharing behaviors [66], and thus, social media interactivity is somehow the basis for the dissemination of tourism crises on social media [67]. Because of social media interactivity, people can learn about a destination in the dissemination of tourism crises, so that they can generate a relatively complete view of the destination, which can lead to a more complete TDI. Thus, an additional hypothesis is proposed as the following:

Hypothesis 3b (H3b).

Social media interactivity positively influences organic cognitive image.

Meanwhile, social media interactivity can increase the quality of information dissemination and reduce information distortion [68]. This can prevent the generation of rumors regarding tourism crises, ensuring the authenticity and objectivity of the released information, thereby enhancing people’s trust in the tourism destination. Thus, a further hypothesis is proposed as the following:

Hypothesis 3c (H3c).

Social media interactivity positively influences organic affective image.

3.4. Organic Cognitive Image, Organic Affective Image, and Organic Overall Image

There have been many studies that have demonstrated the relationship between cognitive image, affective image, and overall image [40,41,69], where cognitive image positively influences affective image, and both cognitive image and affective image positively influence the overall image.

When a tourism crisis is disseminated on social media, the information related to the organic cognitive image suggests whether people’s impressions of the destination are profound. When the organic cognitive image is rated higher, people may mentally construct higher psychological expectations of the destination, which are manifested in positive emotions towards the destination, such as the expected pleasant and relaxing experience, thus improving people’s organic affective image. Organic cognitive image and organic affective image are two aspects of the organic image, either of which can be improved to promote the organic overall image. Thus, additional hypotheses are proposed as the following:

Hypothesis 4 (H4).

Organic cognitive image positively influences organic affective image.

Hypothesis 5 (H5).

Organic cognitive image positively influences organic overall image.

Hypothesis 6 (H6).

Organic affective image positively influences organic overall image.

The purpose of this study is to verify what factors exist in tourism crisis dissemination on social media that positively influence the organic image of tourism destinations. Combining the above hypotheses, in an information dimension, we need to verify the path relationships between information transparency and organic cognitive image, emotional sympathy, and organic affective image; in an information subject dimension, we need to verify the path relationships between the subject authority and organic cognitive image, social connection, and organic affective image; and in an information environment dimension, we need to verify the path relationships between social media stability and organic cognitive image, social media interactivity and organic cognitive image, social media interactivity, and organic affective image. The theoretical framework and hypothesis development pertaining to this are shown in Figure 3.

4. Methods

4.1. Case Introduction

Tourism crises come in various types, and different dissemination processes of tourism crises may exhibit distinct characteristics. We take into consideration that tourism crises caused by management mistakes can particularly damage TDI [9], as the responsibility for crises caused by management mistakes is often more evident compared to events like natural disasters or terrorist attacks. People tend to pay more attention to and condemn these more avoidable events [70]. Thus, we choose to focus on tourism crises caused by management mistakes as our research object, as this holds greater practical significance for crisis management and TDI restoration. Crowd crush is a typical tourism crisis resulting from management mistakes, where excessive crowd density leads to a gradual collapse of the crowd due to individuals pushing against each other, resulting in individuals being crushed or suffocated by the weight of others [71]. This event often occurs during large gatherings, festivals, and celebrations at popular tourism destinations. Its occurrence is often associated with human factors related to DMOs, such as improper personnel scheduling, inadequate venue design, and insufficient safety facilities [72]. Thus, crowd crush represents a typical tourism crisis resulting from management mistakes. Subsequently, we chose the Seoul Halloween crowd crush, which was widely spread across social media, as the case of this study.

The Seoul Halloween crowd crush occurred in Itaewon, South Korea. Located at the eastern foot of Namsan in Yongsan-gu, Seoul, Itaewon is a multicultural business district and one of Seoul’s most popular neighborhoods, known for its nightlife and trendy restaurants [73]. On 29 October 2022, a massive stampede occurred in Itaewon. On that night, a Halloween party was held in Itaewon, and the number of people gathered in the neighborhood was estimated to be around 100,000. Most of the casualties were young people who came to the party, with more than 150 confirmed deaths and more than 100 injuries [19,74]. The event sparked heated discussions among netizens, especially on Weibo, China’s largest social media platform, generating numerous topics with over 100 million views. This indicates that the event underwent extensive dissemination on social media. Additionally, for the destination itself, this is also a very typical tourism crisis. Thus, taking these factors into account, we selected this event as the case for investigation to better understand the impact of tourism crisis dissemination on people.

4.2. Research Methods

There were several relationships between variables involved in this study that were not suitable for direct measurement. Considering that well-established measurement tools for these variables have emerged from existing studies, this study chose to test the hypotheses mentioned above using structural equation modeling (SEM) based on maximum likelihood estimation.

4.3. Survey Questionnaire Design

The questionnaire for this study consisted of three sections. The first section was used to determine the reasonableness of the research participants and to ensure that participants were aware of the Seoul Halloween crowd crush but had not developed the idea of proactively traveling to Itaewon to ensure that their interpreted TDI was their organic image. The second section was used to investigate the demographic characteristics of participants, which included three questions on gender, age, and educational attainment. The third section was used for the measurement model for SEM. This section was developed using a 5-point Likert scale, with each measurement item using a scale of 1 to 5 to quantify people’s level of agreement. Specific items and their sources are detailed in

Table 1. Measurement items.

Latent Variable	Number	Measurement Item	Source of Item
Information transparency (TRAN)	TRAN1	Various publicly available information about the event can be found on Weibo	Agozie et al. [75], Lin et al. [54]
	TRAN2	Information about the event viewed on Weibo is well documented
	TRAN3	Information on how the event was handled at the scene will continue to be followed up
	TRAN4	Information about the event learned through social media explains in detail how it happened
Emotional sympathy (ES)	ES1	I’m sensitive to hot news of the event	Mehrabian et al. [76], An et al. [77]
	ES2	I can empathize with the feelings of strangers who have experienced the event
	ES3	I have great sympathy for the victims in the event
	ES4	Seeing news reports related to the event reminds me of bad things I’ve experienced in the past
Subjective authority (SA)	SA1	I believe the information released by media outlets during the dissemination of the event as it is considered reliable	Kitsios et al. [78]
	SA2	I trust the information released by media outlets during the dissemination of the event because it is helpful to me
	SA3	News reports about the event from various media outlets are generally truthful
	SA4	Various media outlets are important channels for me to obtain news related to the event
Social connection (SC)	SC1	In the dissemination of the event, I established close interactive relationships with others discussing the event	Chiu et al. [79]
	SC2	In the dissemination of the event, I would spend a significant amount of time interacting with others and discussing the event
	SC3	I am aware of the primary news media outlets that played a significant role in the dissemination of the event
	SC4	I’ll discuss the event with someone I know
Social media stability (SMS)	SMS1	Getting into the discussion of the event was convenient and stable for me	Al-Adwan et al. [80]
	SMS2	Breach information about the event can be dealt with quickly
	SMS3	Social media can provide appropriate online assistance and help
Social media interactivity (SMI)	SMI1	My social media friends prompted me to pay attention to this event	Chang [81]
	SMI2	A lot of people around me are following the event and talking about it
	SMI3	Joining in the discussion of the event will facilitate my interaction with others
	SMI4	The atmosphere of information sharing on social media prompted me to get involved
Organic cognitive image (OCI)	OCI1	According to the impression you already have, Itaewon’s tourism resource elements are attractive	Echtner et al. [82], Kim and Yoon [41]
	OCI2	According to the impression you already have, Itaewon is convenient for traveling and booking accommodation
	OCI3	According to your impression, Itaewon has a lot of entertainment facilities
	OCI4	According to the impression you already have, Itaewon welcomes you to visit
Organic affective image (OAI)	OAI1	If you travel to Itaewon, do you think your mood is unpleasant or pleasant	Russell et al. [83], Baloglu and McCleary [40]
	OAI2	If you travel to Itaewon, do you think your mood is lethargic or uplifting
	OAI3	If you travel to Itaewon, do you think your mood is melancholic or exciting
	OAI4	If you travel to Itaewon, do you find your mood is depressing or relaxing
Organic overall image (OOI)	OOI	How would you rate the image of Itaewon as a tourist destination	Baloglu and McCleary [40]

.

4.4. Data Collection

Weibo is the largest social media platform in China, with topics related to the Seoul Halloween crowd crush accumulating over 1 billion clicks. Therefore, we chose to recruit participants from Weibo. Initially, we posted recruitment notices through specific accounts, detailing the purpose of the study, the events being investigated, and the methods of the survey. These notices were regularly updated, and collaborations were established with influential users on Weibo to expand the reach of the survey. Finally, we utilized Weibo’s direct messaging feature to address participants’ inquiries and concerns, thereby enhancing the completion rate and quality of the survey [84].

In the questionnaire introduction, we presented the Seoul Halloween crowd crush to help participants recall the event while answering the questionnaire. To ensure that the measured TDI corresponded to the organic image, a screening question was included at the beginning of the questionnaire: Have you previously visited Itaewon or proactively sought information about related tourism? If answered “Yes”, this indicated that the participant’s TDI at this point was not an organic image, and thus, they were unable to complete the questionnaire. Additionally, a question was added to verify whether participants were answering sincerely to facilitate the subsequent selection of valid questionnaires. The questionnaire was distributed online via Weibo from June 2023 to September 2023. Throughout the survey process, it was ensured that all participants fully understood whether anonymity was guaranteed, the purpose of the study, how their data would be used, and if there were any associated risks before completing the questionnaire. Ultimately, 326 questionnaires were collected, with 247 deemed valid, meeting the basic sample size requirements for SEM [85,86].

5. Results

5.1. Participant Characteristics

The research results show that the gender ratio among the participants is fairly balanced, and the age distribution aligns with the characteristics of internet user groups. The majority fall within the 18–40 age range, and their educational attainment is predominantly at the undergraduate level and above, further ensuring the overall quality of the data. To test whether there was an impact of different population characteristics on the dependent variable, we used one-way analysis of variance (ANOVA) to test whether there was a significant difference between the different populations in terms of organic overall image. The results show that all p-values are above 0.05 (

Table 2. ANOVA of participant characteristics and organic overall image.

Participant Characteristic	Category	Frequency	Proportion	F	p-Value
Gender	Male	131	53.04%	0.712	0.403
	Female	116	46.96%
Age	18 to 30	162	65.59%	1.184	0.315
	31 to 40	75	30.36%
	41 to 50	10	4.05%
Educational attainment	Undergraduate and above	148	59.92%	0.556	0.459
	High school	55	22.27%
	Junior high school	40	16.19%
	Primary school and below	4	1.62%

), indicating that there was no significant difference in people’s organic overall image in terms of gender, age, and education, which could indicate that the participants do not have extreme attitudinal bias and are representative.

5.2. Measurement Model

5.2.1. Reliability Analysis

Reliability analysis can be understood as the degree to which a measurement item produces consistent results if the measurement is repeated several times. Cronbach’s alpha is one of the most commonly used indicators in reliability analysis [87]. It takes a value ranging from 0 to 1, where a value closer to 1 indicates a higher internal consistency of the measurement items. It is generally considered that 0.6 is the cut-off line, higher than 0.6 is the acceptable range, and lower than 0.6 indicates that the measurement item’s reliability fails to be qualified. The results of the reliability analysis are shown in

Table 3. Reliability analysis.

Latent Variable	Cronbach’s Alpha	Overall Cronbach’s Alpha
TRAN	0.735	0.732
ES	0.753
SA	0.742
SC	0.720
SMS	0.710
SMI	0.737
OCI	0.737
OAI	0.775
OOI	0.719

.

According to the results, it can be seen that Cronbach’s alpha of each variable and the whole are above 0.7, which indicates that the internal consistency of the research data is high and meets the requirements for further analysis.

5.2.2. Validity Analysis

Validity analysis is used to examine whether measurement items can effectively express conceptual information about latent variables. The validity analysis in this study was divided into three steps. Firstly, confirmatory factor analysis (CFA) was used to determine whether the measurement items were reasonable and whether there was any overlap between items. Second, convergent validity was determined by the average variance extracted (AVE) and composite reliability (CR). Third, discriminant validity was determined by the Pearson correlation coefficient between variables and the square root of the AVE [88].

The Kaiser–Meyer–Olkin (KMO) test and Bartlett’s test of sphericity were used to determine whether the data met the requirements for CFA. Bartlett’s test of sphericity is used to test whether there are correlations among the variables in data, and if the significance of the result is less than 0.05, this means that there are correlations among the variables and it is suitable for CFA. A KMO value is used to assess the degree of correlation among variables, with a value range of between 0 and 1, and it is usually considered to be greater than 0.6, which meets the basic requirements of factor analysis. The results of the analysis are shown in

Table 4. KMO test and Bartlett’s test of sphericity.

KMO	Bartlett’s Test of Sphericity
0.741	Approximate Chi-square	10,777.358
	Degrees of Freedom	496
	p	0.000

.

The KMO value was greater than 0.7 and the p-value was less than 0.05 in Bartlett’s test of sphericity, so the data from this study met the requirements of CFA. It is worth noting that for the organic overall image, we referred to the study by Baloglu and McCleary, which used a single measurement item [40]. Petrescu mentions that for measuring more structurally specific issues, or for assessing concepts that are simple and lack most of the complexity of mental structure, such as liking a particular purchase, the use of a single item is justified. This is because redundant items at this point can influence people’s intuitive judgments and lead to greater error [89,90]. In SEM, if a latent variable has only one measurement item, this means that the item reflects 100% of its underlying trait construct [91]. So, if the only measurement item in the organic overall image can measure the latent variable with completeness, there is no need to conduct additional CFA on this variable. CFA was completed through Amos to obtain the standardized factor loading coefficients (

Table 5. Standardized factor loading coefficients.

Measurement Item	Factor Loading Coefficient	Std. Error	z	p	Standardized Factor Loading Coefficient
TRAN1	1.000	–	–	–	0.793
TRAN2	1.345	0.146	9.235	0.000	0.708
TRAN3	1.185	0.134	8.830	0.000	0.870
TRAN4	1.642	0.164	10.041	0.000	0.797
ES1	1.000	–	–	–	0.842
ES2	0.978	0.051	18.993	0.000	0.728
ES3	1.323	0.147	8.983	0.000	0.844
ES4	0.932	0.071	13.086	0.000	0.675
SA1	1.000	–	–	–	0.879
SA2	0.931	0.054	17.128	0.000	0.776
SA3	1.053	0.064	16.466	0.000	0.749
SA4	1.154	0.146	7.891	0.000	0.861
SC1	1.000	–	–	–	0.710
SC2	1.262	0.143	8.822	0.000	0.818
SC3	1.845	0.188	9.787	0.000	0.759
SC4	1.765	0.181	9.750	0.000	0.753
SMS1	1.000	–	–	–	0.789
SMS2	0.987	0.095	10.378	0.000	0.816
SMS3	1.115	0.109	10.197	0.000	0.705
SMI1	1.000	–	–	–	0.671
SMI2	1.346	0.091	14.857	0.000	0.842
SMI3	1.202	0.083	14.484	0.000	0.815
SMI4	0.853	0.075	11.435	0.000	0.620
OCI1	1.000	–	–	–	0.717
OCI2	1.109	0.072	15.407	0.000	0.840
OCI3	1.034	0.071	14.511	0.000	0.779
OCI4	1.162	0.083	14.079	0.000	0.753
OAI1	1.000	–	–	–	0.810
OAI2	0.762	0.043	17.733	0.000	0.783
OAI3	0.968	0.049	19.743	0.000	0.849
OAI4	0.976	0.047	20.960	0.000	0.892

).

Factor loading coefficients demonstrate the correlation between latent variables and measured items. If an item shows significance and its standardized factor loading coefficient is greater than 0.6, this indicates a strong correlation, and vice versa, the correlation is weak, and items are not set up properly. The results of the analysis show that the correlation between measurement items and latent variables is strong, and all measurement items can be well explained by latent variables.

Convergent validity refers to whether different measurement items can produce similar measurement results when measuring the same latent variable, that is, the consistency of different items in the same latent variable. Usually, a variable with an AVE greater than 0.5 and a CR greater than 0.7 indicates high convergent validity [88]. According to the results of the analysis (

Table 6. AVE and CR for each variable.

Latent Variable	AVE	CR
TRAN	0.631	0.872
ES	0.602	0.857
SA	0.669	0.890
SC	0.579	0.846
SMS	0.595	0.815
SMI	0.552	0.829
OCI	0.598	0.856
OAI	0.696	0.902

), the AVE and CR of each latent variable met the requirements, which indicates that the research data have a good convergent validity and the measurement items in the latent variables have a good consistency.

Discriminant validity refers to the ability of different latent variables to be clearly distinguished from each other to prevent excessive correlation or overlap among them. The square root of the AVE can indicate the aggregation of variables, and the Pearson correlation coefficient indicates the correlation between variables. If the aggregation of each latent variable is strong (significantly stronger than the absolute value of Pearson correlation coefficients with the other latent variables), this can indicate that the data have a strong discriminant validity [92]. According to

Table 7. Square roots of AVEs and Pearson correlation coefficients.

Latent Variable	TRAN	ES	SA	SC	SMS	SMI	OCI	OAI
TRAN	0.794
ES	0.489	0.776
SA	0.382	0.284	0.818
SC	0.489	0.548	0.131	0.761
SMS	0.401	0.211	0.529	0.323	0.771
SMI	0.467	0.508	0.225	0.604	0.620	0.743
OCI	0.132	0.285	0.249	0.211	0.388	0.239	0.773
OAI	0.170	0.135	0.011	0.087	0.006	0.027	0.267	0.834

, the diagonal numbers are square roots of AVEs, and the rest are absolute values of Pearson correlation coefficients. The square root of the AVE for each latent variable is greater than the Pearson correlation coefficients with the other latent variables, so this means that the data have sufficient discriminant validity.

5.2.3. Model Fit Indices

It is usually considered that a good-quality fit of SEM to the observed data needs to satisfy the following requirements of the commonly used model fit indices: a chi-square divided by the degrees of freedom χ²/df that is less than 3, a goodness-of-fit index (GFI) that is greater than 0.9, a root mean square error of approximation (RMSEA) that is less than 0.1, and a root mean square residual (RMR) that is less than 0.05. This indicates that a model can explain the observed data better [93]. All model fit indices are shown in

Table 8. Model fit indices.

Model Fit Index	χ2/df	GFI	RMSEA	RMR
Standard of Judgment	<3	>0.9	<0.10	<0.05
Results	1.687	0.946	0.054	0.048

, and thus, the fit of the model can be considered as good.

5.3. Structural Model

Path analysis based on maximum likelihood estimation yielded the results of the hypothesis testing for this study, as shown in

Table 9. Results of hypothesis testing.

Hypothesis	Relationship	Path Coefficient	p-Value	Result
H1a	TRAN→OCI	0.165	**	Support
H1b	ES→OAI	−0.256	**	Reject
H2a	SA→OCI	0.343	***	Support
H2b	SC→OAI	0.289	*	Support
H3a	SMS→OCI	0.011	0.905	Reject
H3b	SMI→OCI	0.284	**	Support
H3c	SMI→OAI	0.248	**	Support
H4	OCI→OAI	0.267	***	Support
H5	OCI→OOI	0.716	***	Support
H6	OAI→OOI	0.437	***	Support

Note: *** means p < 0.001, ** means 0.001 < p < 0.01, * means 0.01 < p < 0.05.

.

6. Discussion

Before discussing the impact of the independent variables on TDI in the model, it is worth noting that H4, H5, and H6 are all supported. This indicates that the measurement of TDI in this study conforms to the theoretical structure established in previous research [40], thus demonstrating the validity of the model. Only with this foundation in place does the study of the influence of other variables on TDI hold theoretical significance.

H1a is accepted. This suggests that adequate, accurate, and timely information provided by a destination after the outbreak of a tourism crisis [94] helps to calm people’s fears about the crisis, increases people’s knowledge about the destination, and reduces uncertainty about the crisis [95], thereby positively influencing the cognitive image of the destination. Thus, after the outbreak of tourism crises, DMOs should not deliberately delay the announcement and follow-up of the truth of the event in the news, not to mention deliberately concealing the information related to the event. Instead, they need to disclose information related to the event promptly, so that the public can learn about the cause of the event and the follow-up measures the first time they consume information about it. The more detailed and complete the information, the more people will be able to recognize the comprehensive and complete information about the destination, so that people will have more trust in DMOs [96].

H1b is not accepted, and emotional sympathy negatively influences organic affective image. This suggests that the emotions of empathy, concern, and anger that that people feel as a result of tourism crisis dissemination do not lead to emotional belonging or identification with the destination [97]. DMOs should try to avoid posting content that could trigger strong emotional reactions in crisis situations after a tourism crisis has erupted. This includes avoiding words or images that inspire emotions such as anger, fear, and sadness [15]. Instead, events or situations should be described in a more objective, neutral, and dispassionate manner to reduce emotional charge. So, instead of relying on emotional representations, information publishers can rely on facts and figures to provide objective, accurate, and credible information, which helps the audience to better understand the crisis without being overly influenced by emotional sympathy and exerting emotional condemnation of the destination.

H2a is accepted. This suggests that in tourism crisis dissemination, when information subjects have a high level of authority, the information they release is more likely to improve people’s opinions about tourism crises. People may thus be more willing to learn about the context and reasons for the occurrence of a tourism crisis, leading to a comprehensive understanding of the destination, which contributes to a more positive cognitive image [59]. Thus, after the outbreak of tourism crises, DMOs should standardize the release channels of information to reduce the dissemination of information by unofficial media. Meanwhile, DMOs should also strengthen their cooperation with experts and other opinion leaders with a certain degree of influence and introduce professional third-party viewpoints, to enhance the credibility of the main subject in information dissemination [98].

H2b is accepted. This suggests that for individuals, behaviors that can enhance social connectedness such as interactions on social media, support from friends, and participation in groups that receive tourism crisis dissemination can promote a more positive affective image [99]. Thus, following a tourism crisis, DMOs can proactively participate in social media platforms or through press conferences to engage in discussions with the public about tourism crises, which involves responding to people’s questions, concerns, and feedback to facilitate the creation of stronger social connections in the process [100].

H3a is not accepted and there is no significant influence relationship between social media stability and organic cognitive image. This can be interpreted as a low sensitivity to the information environment stability in the information ecology during tourism crisis dissemination.

H3b is accepted. This suggests that the information environment in which information subjects can interact more may deepen their understanding of the destination, resulting in a more positive cognitive image. This is because greater interactivity represents greater access to a variety of sources of information for the public, leading to a more comprehensive understanding of the destination [101]. H3c is accepted. This suggests that in an information environment with a more interactive atmosphere, people can establish more emotional connections with destinations in tourism crisis dissemination [102], which can help promote a more positive affective image. Thus, instead of blocking information and controlling its dissemination, DMOs need to be open to the discussion of crisis events. In the course of people’s concern about a crisis, they can provide content related to the destination promotion and then arouse people’s interest and increase their motivation to participate in the discussion [103].

Several studies have been conducted to scientifically analyze the Seoul Halloween crowd crush, and these studies have provided ample evidence for DMOs on how to manage tourism crises. For example, Kyoo-Man reviewed the handling of such events through qualitative analyses, suggesting that all stakeholders need to supplement ordinary event contingency planning with national special event contingency planning [18]. This is helpful for DMOs to establish effective tourism crisis prevention mechanisms. Joo investigated the perceived risk associated with crowd crush using this event as an example and demonstrated that perceived risk significantly influences tourists’ anticipated emotions and visit intentions [19]. Mao made recommendations for improving emergency management policies from three perspectives: public opinion, technology, and management [17]. These studies have focused mainly on policy systems and tourist behaviors but also noted the importance of information dissemination. Just as Joo suggests that DMOs can publicize the improved safety of Itaewon by working with influencers in social networks, this has the potential to raise the social awareness associated with visiting Itaewon [19]. This suggests that the restoration of TDI after a crisis event is very important, while restoration requires managing the dissemination of the event. This study specifically summarizes the factors that reduce the harm of tourism crises in terms of information dissemination. In today’s era of faster information dissemination, DMOs need to be more proactively involved in the dissemination of tourism crises. The countermeasures proposed in this study can precisely provide a concrete course of action for the use of social media by DMOs to minimize the impact of the dissemination of tourism crises on the long-term shaping of TDI and to promote the sustainable development of TDI.

7. Conclusions

In an information dimension, information transparency positively influences organic cognitive image while emotional sympathy negatively influences organic affective image. Thus, after the outbreak of a tourism crisis, DMOs should avoid deliberately concealing information about the event and should disclose it promptly, so that the public can be the first to understand the cause of the event and the subsequent measures taken to deal with it. Meanwhile, they should avoid publishing content that may trigger strong emotional reactions. Facts and data should be used as the basis for describing events with objective, neutral, and calm expressions, to reduce people’s emotional impulses to the information content, help them to understand the crisis more rationally, and slow down their emotional condemnation of the destination.

In an information subject dimension, a subject authority positively influences organic cognitive image and social connection also positively influences organic affective image. Thus, DMOs should standardize information dissemination channels, strengthen cooperation with opinion leaders with a certain degree of influence, such as crisis management experts, and introduce professional third-party perspectives to enhance their credibility. Meanwhile, DMOs should proactively participate in the discussion of tourism crises through social media interaction or press conference organization, respond to people’s questions, concerns, and feedback, and promote the establishment of closer social ties in the process.

In an information environment dimension, social media interactivity positively influences organic cognitive image and organic affective image, but social media stability does not significantly influence organic cognitive image. Thus, DMOs should adopt an open attitude towards the discussion of crisis events, then provide content relevant to the promotion of the destination in the context of people’s interest in the event, to stimulate people’s interest in the destination and increase their motivation to participate in the discussion.

In today’s era of social media for all, online public opinion triggered by tourism crises is inevitable. Thus, in the face of tourism crisis dissemination, we should not just closely suppress and control it, but should hold a dialectical point of view to deal with it. We should make use of the factors in tourism crisis dissemination that are favorable to TDI to reduce the harm of the tourism crisis, so that tourism crises can coexist more amicably with the development of the tourism industry, and promote the sustainable development of TDI.

This study proposes new perspectives on the dissemination of tourism crises, yet there are limitations awaiting exploration. First, our sample was drawn from Chinese social media, with participants being predominantly Chinese. However, different countries, cultures, and their use of different social media platforms may influence people’s perceptions of tourism crises. Therefore, it is necessary to select more diverse samples in future research. Second, as mentioned earlier, tourism crises come in various types. While our analysis results are more persuasive for crises caused by human management mistakes, they may not fully apply to crises caused by natural disasters, terrorist attacks, political factors, and so forth. Hence, in our next steps, we aim to select a broader range of tourism crises as the research object and seek universal patterns among them. Third, in this study, we primarily employed SEM as the analytical method, with data primarily sourced from self-reports of respondents. Currently, studies on social media often consider online texts as important data sources. Thus, we hope to validate our current conclusions through online texts in future research. Finally, information ecology is a complex concept, and this study only mainly has the structure of information ecology. To fully understand the dissemination of tourism crises, processes, and temporal variations in information flow, information ecology could be fully taken into account in subsequent studies.