Research on the Impact of Media Credibility on Risk Perception of COVID-19 and the Sustainable Travel Intention of Chinese Residents Based on an Extended TPB Model in the Post-Pandemic Context

Abstract

This study is the first to examine the impact of media credibility on the sustainable travel intention of Chinese residents in the post-pandemic context. Specifically, the mechanisms by which media credibility influenced the sustainable travel intentions of Chinese residents through risk perception are studied. This study developed an extended theory of planned behavior (TPB) model and used a structural equation model (SEM) to analyze the 1219 valid samples received from online questionnaires. The results revealed that media credibility has a negative impact on risk perception of COVID-19 in the post-pandemic context. This suggested that trusted media, messages, and information sources can reduce the risk perception of COVID-19 when individuals contemplate travel. Risk perception negatively affects subjective norms, attitudes, and perceived behavioral control, while these three variables positively influence sustainable travel intention. Significantly, subjective norms have a stronger impact on the sustainable travel intention of Chinese residents than the remaining variables, demonstrating that, in a collective society, an individual’s intention to travel is more susceptible to influence by government sanctions as well as the unsupported opinions of their family and friends. This study makes up for the lack of focus on the media in sustainable tourism research and provides novel insights for future studies.

1. Introduction

In the last two years, China entered the post-pandemic era, which means that the COVID-19 pandemic is largely under control in China, but has not subsided and may break out in a small scale or even large-scale event at any time [1,2,3]. However, studies have shown that COVID-19 has triggered a wide range of travel fears [4,5,6,7,8]. Moreover, the continuous variation and mutation of COVID-19 reflects its capricious nature, with the internet being flooded with mass amounts of false information about the COVID-19 pandemic [9,10,11], which may also lead people to panic, fear, or hesitation to travel. Therefore, it is significant to explore which factors influence the sustainable travel intention of Chinese residents in the post-pandemic context with the theory of planned behavior (TPB).

In the Internet context, it is crucial to consider the media factor as a variable in predicting the risk perception of COVID-19 and sustainable travel intention. The majority of people actually perceive the risk of COVID-19 through information and communications mediums rather than personal direct experience, so the media to a large extent has become the main means to spread, amplify, and stigmatize such risk, and thus change the behavior of the public [12,13]. For example, the information released by the official and authoritative media can—to a large extent—reduce the risk perception of COVID-19, and thus influence an individual’s travel decision-making [14]. In addition, perceived risk is one of the most pivotal factors that tourists should consider when deciding to travel, and thus can highly affect an individual’s attitude and behavior [15,16]. An example from Parady et al. [17] demonstrates this view. The importance of tourist risk perception has been well discussed; however, studies on risk concerned with major public emergencies such as COVID-19 in the Chinese context are scarce. Therefore, this study introduces risk perception as a variable into the TPB model to further examine its impact on tourist sustainable travel intention.

TPB has been widely used in different fields to explicate human behavior such as consumer behavior, health behavior, tourism behavior, and network behavior [18,19,20,21,22]. Such a theory illuminates in detail how the behavioral intention of an individual is affected by subjective norms, attitudes and perceived behavioral control. Several studies have applied the TPB model to explicate behavioral intention in the context of the COVID-19 pandemic, such as food security behaviors [23], the inclination to online consumption [24], and the intention to wear a face mask [25]. In addition, studies concerning travel behavioral intentions using the TPB model in the context of the pandemic were conducted in different countries, such as South Korea [26], Indonesia [27], India [28], and Spain [29]. These studies introduced other variables within the TPB model on the basis of certain research situations, mainly encompassing knowledge, non-pharmaceutical, and facility risk. These variables thereby improve the capability of the TPB to explain human behavior in a certain context, although the current study did not adopt these variables given they were either inappropriate or not the main variables that would theoretically influence the sustainable travel intentions of Chinese residents in the post-pandemic context.

To our knowledge, there has not been a study previously that integrated media credibility as a variable into the TPB model to explore the mechanisms by which media credibility influenced the risk perception of COVID-19 and the sustainable travel intentions of residents in China in the post-pandemic context. Based on the above, the objectives of the current study are: (1) to unmask the decision-making mechanism of individuals’ sustainable intention to travel by applying the TPB theoretical framework, and (2) to explore the influences of two COVID-19 pandemic-related variables (i.e., media credibility and risk perception of COVID-19) on individuals’ decision-making processes.

2. Literature Review and Hypotheses Formulation

2.1. Key Concepts in the Study

2.1.1. Media Credibility

Media credibility refers to the degree to which the audience perceives that the information disseminator and their content are to be trusted [30], and is a subjective cognition of users. Media credibility is the core variable used to evaluate the audience’s perception of media and information in communication, and the scope of media credibility research has been expanded from information sources to message content and communication channels [31]. The credibility of sources, media, and messages is used to measure media credibility. Source credibility is defined as the “judgments made by a perceiver concerning the believability of a communicator” [32]; Yan and Wen [14] illustrated that medical experts, opinion leaders, authoritative media and the government are communication sources to have high credibility during the pandemic. Jenkins et al. [33] considered that media credibility focuses on the credibility of information transmission channels, while message credibility tends to focus on the information itself [30].

2.1.2. Risk Perception

As a theoretical concept in cognitive psychology, risk perception has been used in sociology, management, tourism, and other disciplines in order to deeply understand people’s psychological and behavioral responses. Parady et al. [17] illustrated that tourists will be accompanied by risks in the process of tourism, and risk perception is the possibility that tourists believe that danger may actually occur to them, which is subjective rather than objective. This means that tourists make subjective judgments based on potential negative results (e.g., possible loss) or adverse consequences (e.g., exposure risks). As such, the unsafe factors or risks wrought by public health events, such as COVID-19, on tourism have an important impact on the behavioral intention of tourists [15]. Thus, risk perception has been as a key factor affecting travel decision-making.

2.2. Development of the Research Hypotheses

2.2.1. The Relationships between Media Credibility, Risk Perception and Behavioral Intention

The social amplification of risk framework (SARF) provides a theoretical perspective for exploring the impact of media use on audience behavior. SARF demonstrates that risk amplification is triggered by risk signals or symbols carried in informational exchanges through two specific paths: strengthening or weakening risk signals, or the filtering of redundant signals according to risk attributes and their importance [12]. Studies on risk communication have indicated that tourist risk perception is driven by risk information [34]. The mass media, especially the online media, represent an important subject of risk communication which affects personal subjective risk perception. Gong and Fan [13] illustrated that tourists cannot make travel decisions without eliciting and processing risk information in the context of the COVID-19 pandemic, and tourist media activities regarding the acquisition of pandemic information affect their formation of perceived risks. Li and Ito [35] further demonstrated that information sources and messages with authentic, credible and reliable properties reduced tourist risk perception during the pandemic. The high trustworthiness of the media will reduce the perceived travel risk associated with earthquakes, floods, terrorist attacks and other unexpected emergencies [36]. It can also posit that media credibility is an important antecedent variable of perceived risk, and has a direct impact on risk perception. Based on the above statement, the following hypotheses were proposed:

Hypotheses 1a (H1a).

Media credibility has a negative impact on tourists’ risk perception in China.

Hypotheses 1b (H1b).

Media credibility positively influences tourists’ behavioral intention in China.

2.2.2. The Relationships between Risk Perception, Subjective Norm, Attitude, Perceived Behavior Control, and Behavior Attention

Existing studies have illustrated that major public health events, such as COVID-19, affect public risk perception, which is a key variable to explaining public decision-making behavior [26,27,28]. Nevertheless, risk perception has an important impact on individual behavior and attitude. When faced with the same risk event, individuals with a high level of risk perception tend to take preventative or avoidance measures to reduce their losses. In this regard, numerous studies conducted in the tourism field have demonstrated that risk perception has a significant impact on TPB variables. For example, Yang et al. [4] contended that COVID-19 has triggered a wide range of travel fears, and tourists will reduce or avoid their travel behavior during the crisis. This result was also confirmed by the work of Kozak [37]. Tapsall et al. [38] further verified that there is a negative correlation between tourists’ perceived risks and their attitudes present in cruise-based tourism. Liu and Shi [39] further demonstrated that sports-based tourism consumers harbor a high level of risk perception, which plays a negative role in restricting their attitudes and intentions against the background of the COVID-19 pandemic. Therefore, it can be predicted that risk perception significantly affects tourists’ attitudes and intentions. In addition, risk perception also has negative effects on tourists’ subjective norms and perceived behavioral control [26]. When tourists believe that there are risks in tourism during the pandemic, they will doubt whether the mainstream public opinion, or that of their relatives and friends, support their tourism, and whether their self-control ability can reduce the obstacles caused by risks. Based on the above statement, the following hypotheses were proposed:

Hypotheses 2a (H2a).

Risk perception has a negative effect on tourists’ attitude.

Hypotheses 2b (H2b).

Risk perception has a negative effect on the subjective norms of tourists.

Hypotheses 2c (H2c).

Risk perception has a negative effect on the perceived behavioral control of tourists.

Hypotheses 2d (H2d).

Risk perception has a negative effect on the behavioral intention of tourists.

2.2.3. The Relationships between Subjective Norm, Attitude, Perceived Behavior Control, and Behavior Attention

Abundant studies that have applied the TPB model have demonstrated that attitudes, subjective norms and perceived behavioral control respectively positively influence travel intentions. For example, to explain the tourists’ behavioral intention regarding coastal tourism in Thailand, Panwanitdumrong and Chen [40] highlighted that tourists’ subjective norms, attitudes, and behavioral control occupy a positive leading role for tourists to construct environmentally friendly behavior. Wu [41] studied the behavioral intention of citizens in the cities of Changsha, Zhuhzou, and Xiangtan in China toward low-carbon tourism and found that attitude along with perceived behavioral control were the two most significant factors to heavily and positively influence low-carbon tourism motivation while the predictive power of subjective norm on the intention of tourists was relatively limited. Based on the above statement, the following hypotheses were proposed:

Hypotheses 3a (H3a).

Subjective norms positively influence the behavioral intention of tourists in China.

Hypotheses 3b (H3b).

Subjective norms positively affect the attitudes of tourists in China.

Hypotheses 4a (H4a).

Perception behavioral control positively affects tourists’ behavioral intention.

Hypotheses 4b (H4b).

Perception behavioral control positively affects tourists’ attitudes.

Hypotheses 5 (H5).

Attitude has a positive impact on the behavioral intention of tourists.

3. Research Design

3.1. Measuring of the Constructs

The hypothesized research model integrated media credibility and risk perception into TPB that was empirically measured using online survey data collected from tourists concerning the impact of media credibility on behavioral intention through risk perception (See Figure 1). In the current study, the questionnaire items employed maturity scales that have been verified in prior studies, such as that of Ajzen [42], Rahmafitria et al. [34], Pahrudin [43], to guarantee content validity. Furthermore, certain appropriate revisions were made so as to better gauge the touristic behavior of Chinese residents in the post-pandemic context. Moreover, in order to refine the questionnaire items for improved measurement accuracy, 26 participants were interviewed via social media. Finally, a pilot study was conducted to evaluate and refine the measurement items so as to ensure content validity.

The entire questionnaire consisted of three sections: the introduction, the demographic information, and the scales from the extended TPB framework. The first part briefly introduced the aim and content of this study and expressed our gratitude to the participants, while the second part encompassed the nominal scales, mainly including gender, age occupation, monthly income and education, and the number of travelers in the post pandemic context. The final part includes the six measurement variables: media credibility, risk perception, attitude, subjective norm, perceived behavioral control, and behavioral intention. Each item was measured via a five-point Likert scale anchored from strongly disagree (1) through neutral (3) to strongly agree (5).

• Media credibility: three items adopted from Metzger et al. [30] and Yan and Wen’s [14]; Sample items: Information related to the COVID-19 pandemic from medical experts, government officials and COVID-19 infected cases is reliable.
• Risk perception: four items modified from Sánchez-Cañizares et al. [29], Liu et al. [44], and Cori et al. [45]. Sample example: I think it will be easier to be infected by COVID-19 when traveling.
• Attitude: five items derived and slightly modified from Azjen [42], Seong and Hong [46], and Liu and Shi [39]. Sample item: I think it is good for my health to travel in the post-pandemic context.
• Subjective norm: six items derived and modified from Azjen [42], Xu et al. [47], and Harmid and Bano [48], sample item: My relatives support me travelling in the post-pandemic context.
• Perceived behavioral control: five items modified from Harmid and Bano [48] and Pahrudin et al. [43]. Sample item: I am confident that traveling in the post-pandemic context is entirely within my control.
• Sustainable travel intention: four items derived and modified from Pahrudin et al. [43]. Sample item: I will continue to travel in the post-pandemic context.

The whole measurement items in the questionnaire are presented in the Appendix A below.

3.2. Sampling and Subjects

The questionnaire was created through a professional online questionnaire survey platform called Wenjuanxing (https://www.wjx.cn/, accessed on 27 January 2022), which focuses on providing users with a series of services such as powerful and humanized online questionnaire design, data collection, and survey result analysis. To ensure that each participant completed the questionnaire only once, the study gave them an ID that could only be used to answer the questionnaire once. This questionnaire was disseminated at random through social media, such as WeChat, Microblog, and QQ, and it lasted from 27 January to 15 April 2022. A total of 1235 questionnaires were recovered from 34 provinces in China, of which we removed 16 which were incomplete. The full sample data is available on request. Finally, the 1219 remaining valid questionnaires were kept for subsequent data analysis and empirical model testing.

The demographics of the overall sample have been summarized in

Table 1. Demographics of respondents.

Variable	Characteristics	Frequency	Percentage
Gender	Female	528	43.31%
	Male	691	56.69%
Age	18–25	594	48.73%
	26–45	459	37.65%
	46–65	162	13.29%
	Above 65	4	0.33%
Occupation	Student	385	31.58%
	Enterprise staff	328	26.91%
	Public official	288	23.63%
	Freelancer	109	8.94%
	Retiree	25	2.05%
	Other	84	6.89%
Monthly income	Less than 1500 RMB	245	20.10%
	1500–3000 RMB	193	15.83%
	3001–4500 RMB	210	17.23%
	4501–6000 RMB	284	23.30%
	Over 6000 RMB	287	23.54%
Educational level	Higher school education and below	183	15.01%
	Undergraduate	803	65.87%
	Master’s degree and above	233	19.11%

. The findings of this current study highlighted that the majority of the respondents were males (56.69%), while females occupied 43.31%. The age of 48.73% of the respondents was between 18 and 25, tied for the largest proportion, while the age of respondents above 65 accounted for the smallest proportion. Additionally, students (31.58%), enterprise staff (26.91%) and public officials (23.63%) demonstrated greater interest in travelling. The majority of the respondents (46.84%) had an income over 4501 RMB per month. Respondents whose income was between 15,001–1000 RMB accounted for the smallest proportion (15.83%). Moreover, the findings indicated that most of the respondents had undergraduate qualifications (n = 803, 65.87%), followed by those who had postgraduate level qualifications (n = 233, 19.11%).

4. Data Analysis and Empirical Results

Structural equation modeling (SEM) is a multivariate statistical analysis technique that is used to analyze structural relationships, and has been widely used in the social sciences [49,50]. It provides a flexible framework for developing and analyzing complex relationships between multiple variables, and allows researchers to use empirical models to test the effectiveness of a theory. Beran and Violato [51] stressed that its greatest merit might be its capacity to administer measurement error, which is one of the greatest limitations of most studies. In this current study, we employed the SPSS 26.0 and AMOS 24.0 statistical software to analyze the research samples. More specifically, SPSS 26.0 was used to implement the descriptive statistics, reliability analysis, and correlation analysis, while AMOS 24.0 was applied to conduct the confirmatory factor analysis and structural equation model analysis. The results are presented as follows.

4.1. Reliability and Validity Tests

In the current study, Cronbach’s Alpha was used to measure the reliability of the questionnaire.

Table 2. Factor loadings and confidence levels.

Construct	Items	Factor Loadings	t-Value	Cronbach’s a	CompositeReliability	AVE	Mean	SD
MC	MC1	0.837		0.759	0.794	0.566	3.94	0.678
	MC2	0.759	23.152				3.78	0.658
	MC3	0.648	20.788				3.90	0.678
RP	RP1	0.780		0.880	0.882	0.601	2.15	1.051
	RP2	0.827	30.158				1.96	1.086
	RP3	0.815	29.683				1.97	1.113
	RP4	0.666	23.539				2.32	1.168
	RP5	0.777	28.119				2.61	1.148
AD	AD1	0.727		0.848	0.851	0.533	3.96	0.831
	AD2	0.708	22.983				3.79	1.037
	AD3	0.769	24.849				3.99	0.827
	AD4	0.721	23.384				3.97	0.889
	AD5	0.723	23.453				3.85	0.884
SN	SN1	0.697		0.864	0.864	0.515	4.03	0.844
	SN2	0.734	22.928				3.84	0.950
	SN3	0.669	21.070				3.90	0.899
	SN4	0.737	22.988				3.90	0.929
	SN5	0.724	22.626				3.80	0.962
	SN6	0.742	23.139				3.82	0.992
PBC	PBC1	0.745		0.857	0.858	0.547	3.79	0.999
	PBC2	0.744	24.651				3.83	0.914
	PBC3	0.703	23.299				3.92	0.875
	PBC4	0.739	24.490				3.82	0.936
	PBC5	0.764	25.294				3.78	0.975
BI	BI1	0.705		0.856	0.859	0.604	4.03	0.882
	BI2	0.768	24.445				3.79	1.084
	BI3	0.798	25.283				3.81	0.985
	BI4	0.832	26.172				3.51	0.968

Note: DM: Media credibility; MCB: Medium credibility; MDC: Message credibility; SCB: Source credibility; RP: Risk perception; Attitude: AD; SN: Subjective norm; PBC: Perceived behavioral control; BI: Behavioral intention; SD; standard deviation.

presents that the value of the reliability of the six constructs are 0.759, 0.880, 0.848, 0.864, 0.857, 0.856, respectively, which are all greater than 0.7 [52], indicating that the reliability of the data is sufficient for further analysis. Subsequently, confirmatory factor analysis (CFA) was conducted to test convergent validity and discriminant validity (Table 2). Convergent validity can be tested using the factor loadings and average variance extracted (AVE) of the variables. As displayed in Table 2 [53], the values of the factor loadings of the six constructs ranged between 0.648 and 0.837, which were all greater than 0.60 (>0.5), indicating that all of the items had a strong correlation with the corresponding factor. The AVE values were between 0.515 and 0.604 (>0.5), and the values of the composite reliability (CR) were between 0.749 and 0.880 (>0.7). Therefore, the scale of each construct had good convergent validity.

In addition,

Table 3. Correlation coefficient matrix.

Construct	MC	RP	AD	SN	PBC	BI
MC	0.752
RP	−0.385	0.775
AD	0.250	−0.371	0.730
SN	0.261	−0.315	0.519	0.718
PCB	0.271	−0.391	0.511	0.319	0.740
BI	0.475	−0.488	0.622	0.468	0.508	0.777
Mean	3.777	2.203	3.910	3.879	3.828	3.785
SD	0.658	0.915	0.707	0.718	0.751	0.821

Note: Diagonal elements in boldface are the values of average variance extracted. Off-diagonal elements are the correlation coefficients.

presents that the values of the square roots of the AVEs corresponding to the six factors were 0.752, 0.770, 0.730, 0.718, 0.740, 0.777, respectively, and the correlation coefficient between the factors was between 0.250 and 0.622. The minimum value in the square root of AVE was greater than the maximum value of the correlation coefficient between factors of 0.622, verifying that the scale had good discriminant validity.

4.2. Common Method Bias Testing

The sample data in the study was collected through self-reported surveys in a single setting, which is necessary to assess common method variation (CMV). According to Podsakoff et al. [54], this refers to the false common variation between traits caused by the use of the same measurement tool, which is often the case with data that has been measured by self-reported scales. The deviation caused by CMV is called common method bias, which is a systematic error independent of traits and which affects the validity of the measurement. In order to ensure the accuracy of the data, this study therefore utilized Harman’s single factor test, which is a technique used to check common method bias [55]. The test results illustrated that the variance interpretation rate of the first factor was 26.061, which is less than 40% [56]; moreover, the fitting of the single factor model did not meet the standard (see Table 2) in the confirmatory factor analysis. These testified to the fact that there was no significant common method bias in our measurement.

4.3. Hypothetical Research Model Test

4.3.1. Goodness-of-Fit Test

The current study used the maximum likelihood method to assess the goodness-of-fit of the measurement model. The goodness-of-fit criteria for the measurement model should identify that the model sufficiently interprets the empirical data [57].

Table 4. Goodness-of-fit indices for the measurement scales.

Fit Index	Recommended Value	Measurement Model	Structural Model	Sources
$\chi$2/df	<5	2.194	2.383	Bagozzi and Yi [58]
RMSEA	<0.08	0.031	0.034	Steiger and Lind [59]
SRMR	<0.08	0.026	0.048	Maydeu-Olivares et al. [60]
GFI	>0.9	0.958	0.954	Segars and Grover [61]
RFI	>0.9	0.949	0.945	Bentler and Bonett [62]
NFI	>0.9	0.955	0.951	Hu and Bentler [63]
CFI	>0.9	0.975	0.971	Bentler [64]
TLI	>0.9	0.972	0.967	Kenny and McCoach [65]
IFI	>0.9	0.975	0.971	Bentler [64]
PNFI	>0.05	0.846	0.852	Bentler and Bonett [62]; Sahoo [66]
PGFI	>0.05	0.790	0.796	Bentler and Bonett [62]; Sahoo [66]
PCFI	>0.05	0.864	0.870	Bentler and Bonett [62]; Sahoo [66]

Note: RMSEA: root-mean-squared error of approximation; SRMR: standardized root mean square residual; GFI: goodness-of-fit indices; RFI: relative fit index; NFI: normed fit index; CFI: comparative fit index; TLI: Tucker-Lewis index; IFI: incremental fit index; PNFI: parsimony normed fit index; PGFI: parsimony goodness of fit index; PCFI: parsimony comparative fit index.

indicates that the estimates of the whole goodness-of-fit criteria surpassed their respective acceptable values recommended in the previous literature, which suggested that the measurement model in the current study displayed a fairly good fit with the data collected.

4.3.2. Structural Equation Model Testing

The SEM was evaluated through two main measurements: the standardized coefficient (SC) and the t-value (see

Table 5. Summary of hypotheses testing results.

Hypotheses	Paths	Standardized Coefficient	S.E.	t-Value	Supported
H1a	MC → RP	−0.397 ***	0.049	−11.541	Yes
H1b	MC → BI	0.259 ***	0.033	8.399	Yes
H2a	RP → AD	−0.114 ***	0.024	−3.436	Yes
H2b	RP → SN	−0.331 ***	0.030	−9.950	Yes
H2c	RP → PBC	−0.404 ***	0.031	−12.007	Yes
H2d	RP → BI	−0.157 ***	0.025	−4.732	Yes
H3a	SN → AD	0.386 ***	0.027	11.507	Yes
H3b	SN → BI	0.363 ***	0.039	9.446	Yes
H4a	PBC → AD	0.361 ***	0.027	10.489	Yes
H4b	PBC → BI	0.167 ***	0.027	5.119	Yes
H5	AD → BI	0.126 ***	0.026	4.006	Yes

Note: *** p < 0.001.

). The value of the standardized path coefficient between media credibility and risk perception was −0.397 (p < 0.001), indicating that attitude had a significant, direct effect on risk perception at t-value = −11.541, with a negative relationship. The stronger the level of the tourists’ media credibility, the weaker the perceived tourism risk, especially when the COVID-19 pandemic was still not subsiding. Therefore, H1a was supported. Similarly, H1b was verified.

The following four hypotheses were concerned with the effects of risk perception of COVID-19 on TPB. A significant relationship was demonstrated by the effect of physical, psychological, social and financial indicators of perceived risk on one’s attitude towards avoiding COVID-19 risk, with a SC value = −0.114 (p < 0.001) and a t-value = −3.436, indicating a negative direct impact on the attitudes of tourists. Therefore, the stronger the COVID-19 risks perceived by the tourists, the stronger their attitudes would be towards avoiding travelling. The direct relationship between risk perception and subjective norm was negative, with a SC value of −0.331 (p < 0.001) and t-value of −9.950, showing that the stronger the COVID-19 risks perceived by tourists, the weaker the support from their family and significant others (such as family, friends, and relatives) and the media and government to travel. In addition, the perceived risk of COVID-19 has a direct and negative influence on perceived behavioral control, with a path coefficient of −0.404 (p < 0.001) and a t-value of −12.007. This result indicated that the lower an individual’s perceived risk concerning COVID-19 is, the lower their self-control regarding whether or not to travel in the pandemic context. Furthermore, the values for SC (−0.157 ***) and t-value (−4.732) revealed that the risk perception of COVID-19 also has a negative relationship with individuals’ behavioral intentions. This outcome considers that the higher the risk perception of COVID-19, the lower the intention to travel. Therefore, the four hypotheses (H2a, H2b, H2c, and H2d) were all verified.

The SC value between subjective norm and attitude was 0.386 (p < 0.001), indicating that subjective norms directly and positively influence attitudes at t-statistic value = 11.507. Subjective norm also has a significant relationship with intention of behavior, as indicated by the SC value (0.363, p < 0.001) and t-value (9.446). The results suggest that the more the significant people/organizations (friends, family, relatives, the media and government) around an individual support their decision to travel, the stronger their attitude/intention to travel will be (as long as the COVID-19 pandemic has not vanished). H4a and H4b were concerned with the effects of perceived behavioral control on attitude and behavioral intention. The results indicated that perceived behavioral control has a positive impact on attitude (SC value = 0.361 ***, t-value = 10.489) and behavioral intention (SC value = 0.167 ***, t-value = 5.119), respectively. This means that the stronger an individual’s ability to control behavior, the stronger their attitude/intention to travel. H5 is concerned with the impact of attitude and behavioral intention. A path coefficient value (0.126 ***) was gained, which demonstrated that attitude significantly influenced the intention of behavior at t-value = 4.006. Therefore, hypotheses H3a, H3b, H4a, H4b, and H2d were all verified.

5. Conclusions and Discussion

In the post-pandemic context, due to the widespread embeddedness of digital media, social activities such as traveling have more profoundly touched public cognition with the help of discourse expression in media reproduction. This study was the first to focus on the impact of media factors on the risk perception of COVID-19 and the sustainable travel intention of Chinese residents in the post-pandemic context.

Media credibility is a variable that negatively affects the risk perception of COVID-19. This result is consistent with the work of Yang [1] who argued that tourists’ access to sufficient and credible information can reduce their uncertainty caused by COVID-19 pandemic risk, thus reducing tourists’ risk perception. Especially, the more tourists that trust the sources and channels sharing relevant COVID-19 information, the more they agree with them, and the more likely they are to reduce their own risk perception of COVID-19 in the context of travelling. In addition, risk perception plays an intermediary role in the relationship between media credibility and tourists’ behavioral intentions, which confirms the work of Yan and Wen [14]. This illustrates that the significant impact of media credibility on tourist decision-making also reflects the public’s demand for consistency in media information during the pandemic. Moreover, the cognitive risk formed through media credibility is a direct factor that acts on tourists’ behavior, which differs from factual risk [67]. Significantly, recognizing the complexity of media credibility on travel behavior decision-making will assist tourists to objectively judge the potential risks in their tourism, thereby avoiding a fear of travel caused by unclear and asymmetric information, further reducing their own risk perception of COVID-19, and ultimately affecting their own tourism decisions.

Risk perception of COVID-19 negatively influences attitudes, subjective norms, and perceived behavioral control. These findings are consistent with previous studies [68,69] revealing that perceived risks have an influence on TPB variables. This means that the lower a person’s risk perception of the novel coronavirus, the more positive their travel attitude, perceived behavioral control and subjective norms will be, and vice versa. Among those variables, risk perception was found to have the greatest impact on perceived behavioral control followed by subjective norms and attitudes in the context of the COVID-19 pandemic. Touristic behavioral intention is strongly and positively predicted by attitude, perceived behavioral control and subjective norm in the context of traveling during the COVID-19 pandemic in China. In addition, the risk perception of COVID-19 negatively impacted the sustainable travel intentions of Chinese residents in the post-pandemic context. This result has indicated that individuals will perceive the psychological, temporal, monetary and social risks that have been wrought by COVID-19 when traveling. This result is consistent with previous studies [4,37,38,39]. In actuality, COVID-19 has incurred great losses to the Chinese tourism industry during the pandemic. According to the report furnished by the China Tourism Academy, annual Chinese national tourism revenue dropped from 6.63 trillion RMB in 2019 to 2.23 trillion RMB in 2020, and its contribution to the GDP also fell to 2.2% [70], making it the worst year in the history of Chinese tourism [71]. This current study furthermore reveals that COVID-19 still has a significant impact on the Chinese tourism industry in the post-pandemic period. More specifically, the public considers COVID-19 to be a very terrible disease and that they can be easily infected when traveling, which has triggered a wide range of travel fears [4,5,6]. Moreover, the governments in different regions/provinces in China have issued a series of regulations concerning travel in the post-pandemic context [72,73], among which one states that nonlocal tourists must present the negative certificate of a nucleic acid test of either 24 or 48 h old in order to be allowed to enter local tourist attractions. Even locals must display their green health code and green trip code before entering a tourist attraction. This health code and trip code are used to identify people that have been potentially exposed to COVID-19, and the green code signifies that the person is healthy [74,75]. This means that individuals must spend more time and money on traveling in the post-pandemic context.

Significantly, in comparison with the two variables of subjective norm and perceived behavioral control, our study found that subjective norm has the strongest interpretive ability for the sustainable travel intention of Chinese residents in the post-pandemic context. This result is inconsistent with the findings of Liu et al. [44] who found that attitude rather than subjective norm has the strongest potential to predict travel intention, and also consistent with the work of Guo et al. [76] who demonstrated that perceived behavioral control has the strongest impact on behavioral intention. However, our result is in line with the findings from Rahmafitria et al. [34] and Juschten et al. [69]. This shows that a subjective norm may play a controversial role in the context of different research. The reasons for this may be that China is a collectivist society, meaning an individual’s behavioral intention is heavily influenced by social pressures and government regulations. More specifically, the perceptions and norms of the nearest social groups, such as one’s friends, family members, and relatives, including the government, heavily affect the travel intention of an individual. This phenomenon demonstrates that individuals from collectivist societies are more willing to travel if they are supported by positive opinions from the government and their close social circles in the post-pandemic context. Furthermore, a collectivist society, like China, can imperceptibly alter individuals’ views when they are inconsistent with collectivist views.

6. Theoretical and Practical Implications

6.1. Theoretical Implications

Our study presents three major contributions to the theory in particular. First, this study tentatively coupled risk communication with tourism behavior research, and verified the key role of risk perception in the field of tourism behavioral research, especially in the context of potential public health risks. Second, this study sought to enrich the empirical research on the impact of touristic behavior in the post-pandemic context, and plays a certain role in supplementing and improving the theoretical system of travel behavior under potential risk scenarios. Third, this study innovatively introduced media credibility as an antecedent variable, which is a response to the proposal of Nguyen and Nguyen [77] that in the Internet age, the media carries out the functions of exposing, emphasizing, condensing and visualizing risks, yet effective media dissemination mechanisms and the high trustworthiness of the media will reduce risk perception, thereby affecting an individual’s decision-making [13,36].

6.2. Practical Implications

The findings of this study have several practical implications for the tourism sector, government departments, tourism marketing organizations, and other stakeholders. First, the results that subjective norm is the most significant factor influencing sustainable travel intention. The government and social pressure, as the important external factors shaping subjective norms, have become of utmost importance. It is thus necessary for the tourism sector to use advanced technologies, such as artificial intelligence and big data, to monitor public opinions on traveling in the context of the pandemic, thereby correctly guiding the direction of public opinion so as to correct their negative/incorrect views on traveling during the pandemic to reduce social touristic pressure when traveling. Moreover, the government should formulate measures to control the flow of tourists to scenic spots and strictly prevent the gathering of people. In addition, tourism managers should reshape a harmonious and safe image of tourist attraction, reactivate the tourist market, and restore the confidence of tourists.

Second, the perceived risk of COVID-19 had a significant impact on the TBP variables. To effectively reduce the risk perception of COVID-19, tourism enterprises should improve how they release their news and respond to crisis events so as to provide real and objective information to tourists in the beginning. During major public emergencies, mass amounts of fake information and disinformation were disseminated, which magnified people’s perception of risk, so it is important to improve the media literacy of the public so they may actively screen information acquisition channels and rationally identify risk information. The tourism sector must fully understand the informational needs of the public and respond to issues of social concern in a timely manner. Government departments should pay closer attention to the penetration of risk information and effectively act out the role of risk emergency management; the media should thus adhere to the professional spirit of journalism and effectively strengthen the risk communication between the government, the experts, and the public; this is because a rational and open attitude towards risk cognition and information interpretation represents the proper way for the public to deal with public emergencies.

Third, media credibility has a positive impact on sustainable travel intention. In the risk pattern of public emergencies, people are more likely to be in a state of anxiety and panic, so tourism-related sectors should use the media to report the information related to COVID-19 truthfully to avoid any misplaced risk perceptions by the public and subsequent incorrect behavioral decision making. In addition, it is necessary for tourism stakeholders to fully utilize authoritative or official media platforms to share information on the pandemic so as to help the public have a clear understanding of the pandemic, thereby assuaging their panic during a crisis when they are traveling. Furthermore, the government should urgently standardize the role of the media in risk communication, guiding the public to participate and rationally discuss traveling behavior during crisis events.

7. Limitations and Future Research

The results of the current study are of significant value to the development of a better understanding regarding the relationship among media credibility, risk perception, attitude and sustainable travel intention, which has been seldom discussed in the literature on tourism, the media, and the pandemic. However, our study encountered several limitations. First, one traditional limitation is that our study employed a convenient sampling method, which means that the sample data may not represent the entire population in China. Secondly, the current study only incorporates the media credibility into the TPB model as a variable, so there is still a lack of research on the impact of different types of media on behavioral decision-making during crises. Third, many countries worldwide are still severely affected by the Coronavirus, however our study concentrated on Chinese respondents. Therefore, it may be difficult to generalize the results to other countries. Nevertheless, these limitations did not cripple nor impair our findings in any way.

In terms of future studies, new techniques, such as artificial intelligence, are recommended to be used to obtain a larger body of sample data to better explicate the travel behavioral intention of the population influenced by other variables, such as political or financial factors, in the post-pandemic context. Moreover, future research can be conducted to explore the impact of social media platforms, such as WeChat, Facebook (Meta), TikTok, and YouTube, on travel behavioral decision-making in the post-COVID-19 pandemic context. In addition, comparable studies in other Asian countries or countries with different social and cultural backgrounds can be conducted.