Ramification of Covid-19 Perceived Risk on Vacation Intention in Malaysia

COVID-19’s worldwide ramifications are unknown, as the world is currently engulfed within one of the mo st serious last century’s health and humanitarian disasters. Health and safety problems, political changes, economic instability, natural disasters and disease outbreaks threaten tourism. The danger associated with taking a vacation is increased by the global pandemic, not just due to the unknown state of the vacationer in the destination but also because of the possible negative repercussions. Thus, this research fills the gap by analysing the influence of the Theory of Planned Behaviour (TPB) predictors (namely attitude, perceived behavioural control and subjective norm) and perceived risk towards vacation intention and willingness to pay amid local vacationers in Malaysia during the pandemic of COVID-19. The acquired data from 181 completed responses were analysed using SmartPLS software. According to the findings of this study, TPB predictors (attitude, perceived behavioural control, and subjective norm) have a positive influence on vacation intention, perceived risk has a positive influence on TPB predictors (attitude and perceived behavioural control), and vacation intention has a positive influence on willingness to pay. The practical contribution of the research could contribute to the considered actions to foster the necessary decision-making to restore the domestic tourism market.


Introduction
vacationer or visitor's decision-making process. COVID-19 is thus a disastrous phenomenon that is crushing the tourism sector into a tiny one. Travel restriction on vacation from almost every country globally has also contributed to inhibiting tourism activity. According to UNWTO (2020), According to UNWTO (2020), as of September 1, 115 destinations had loosened their travel restrictions, which is an increase of 28 from the number of destinations that had done so on July 19, 2020. This represents 53% of all destinations in the world. Only two of these countries have removed any and all limitations, while the other 113 countries have not done so and still have particular restrictive measures in place. 93 of the most popular tourist locations in the world do not allow tourists beyond their borders (43 per cent of all destinations worldwide). When compared to July 19, 2020, this indicates a decrease of 22 destinations (UNWTO, 2020). Slow containment of viruses postpones global air traffic recovery. Global air traffic is recovering more slowly than expected and will rebound to prepandemic levels by 2024 (Mchugh, 2020). The International Air Transport Association set its prediction back by a year because of the virus's slow containment in the United States and developing countries. The industry is emerging from the shutdowns' depths in April, but the bad news is that any increase is almost invisible (Mchugh, 2020). For that reason, the projected tourism sector will face a disastrous period of more than four years.
The vacation decision entails risk due to the global pandemic, not just because of the uncertainty surrounding the vacationer's situation in the destination but also because of the potential adverse effects (Chang, 2009). Fuchs and Reichel (2006) highlighted that in the tourism sector, there are a few different kinds of risks to consider: risk to one's bodily or mental health (food security, outbreaks, and mishaps are all things to be concerned about.); the risk associated with the equipment (issues with the equipment or the organisation); risk of the psyche (the vacationer is dissatisfied with his or her vacation); monetary risk (unanticipated expenses); societal risk (friends and relatives' attitudes regarding the vacationer have shifted); and risk of running out of time (the vacation may be in futile). While none of these instances is directly tied to COVID-19, vacationers' concerns about health hazards or the chance of contracting an infectious disease have influenced their activities and destination choices in recent years (Chinazzi et al., 2020;Lee et al., 2012).
In the meantime, Hayashi (2020) said service providers seek to compensate in creative ways for lost revenue and increasing expenditures that will reduce the economic effect of . Paying an extra cost or surcharge for a customer's receipt and labelling it "COVID-19" or "sanitation" to prevent the spread of COVID-19 is one method that generates a lot of undesirable negative publicity. Moreover, in reaction to and safeguarding against the spread of COVID-19, Abdallah (2020) stated that many businesses and service providers are steadily charging clients. Many businesses demand surcharges to cover the cost of additional safety precautions, such as personal protective equipment for workers or sanitary supplies. Callahan and Nguyen (2020) concur that many restaurants add a COVID-19 extra to their customers' bills to cover the cost of cleaning materials, personal protection equipment for staff, disposable menus, and other expenses. Zee Krstic (2020) added that a few businesses have already discovered a new way to deal with these overnight changes to their livelihoods; a COVID-19 surcharge is typically included to each bill or order as a percentage. Although some customers may view this as unfair, this additional payment will certainly prevent the company from incurring more pandemic-related expenses or closing its doors entirely.
The TPB (Ajzen, 1991), which is one of the most well-known theoretical frameworks in intention analysis, is the primary focus of the present investigation (Gollwitzer & Sheeran, 2008). The focus of this research is on the varying risk impacts that have an effect on the TPB's antecedent of intention. The model that was developed takes into account the impact that this intention has on willingness to pay (WTP) in order to receive a more substantial benefit from the destination's increased security measures.

Literature Review Perceived Risk
According to a study conducted by Lobb et al (2007), risk perception does not appear to have a direct impact on behavioural intentions. However, it has a negative effect on attitudes, which are the most important factor. Risk perception negatively affects product attitudes. However, the subjective norm, such as social influence sensitivity, appears to be positively linked (in both directions) with trust in food safety information and attitudes. Because perceived risk is an expectation of a future loss, it is likely to have a negative effect on attitudes toward a behaviour Quintal et al. (2010). For instance, the greater the perceived risk of financial loss associated with a purchase, the more negative views regarding the transaction are likely to be. As a result, the hypothesis that follows is put forth H1: The higher the perceived risk of COVID-19, the more unfavourable the perception of taking a vacation during the epidemic.
People's difficulty in behaving, such as making holiday decisions, is measured by perceived behavioural control . The perceived risk is also likely to alter the perception of behavioural control over actions. The greater the perceived risk of financial loss involved with a transaction, for example, the less control people will have over the purchase. Furthermore, people's perceptions of control are likely to suffer when there is more ambiguity regarding the outcome of a decision (Quintal et al., 2010). Quintal et al (2010) said that risk could also affect the PBC over a behaviour. The more significant the perception of potential negative harms resulting from the vacation, the less control the person feels over that action. Thus, the hypothesis that follows is developed: H2: The higher the perceived risk of COVID-19, the lower the perceived behavioural control over the idea of taking a vacation during the pandemic.

Attitude
Previous research has revealed that a person's attitude toward having a vacation at a specific location is a key component in forecasting vacation intentions at that location, validating the TPB's conclusion (Bam & Kunwar, 2019;Hsu & Huang, 2012;Hsu, 2013;Martin et al., 2011;Quintal et al., 2010;Ziadat, 2015). Ziadat (2015) discovered, for example, that tourists' attitudes on visiting and revisiting the destination (in Jordan) had a substantial impact on their desire to return. In parallel, Hsu (2013) found that attitudes toward water leisure activities greatly influence travel intentions to rivers and lakes where those activities are possible. Hsu and Huang (2012) discovered that attitude did play a role in behavioural intention, although it was just a little one. However, Martin et al (2011) found that attitudes toward travelling to a different place for medical care were the most significant predictor of medical tourism intentions. It may be claimed that the strongest beliefs were found in each of these examples. Experience, according to Cooke and Sheeran (2004), is necessary for confirming beliefs. Thus, a positive attitude is anticipated in Ziadat's (2015) study. Similarly, one may assume that the attitude toward medical care is founded on a firm belief in its significance.The following theory is therefore proposed: H3: The higher the vacation intention, the more optimistic the attitude toward the potential of vacationing throughout the COVID-19 pandemic.

Perceived Behavioural Control
Most of the previous research showed that perceived behavioural control over a variety of relevant characteristics substantially influenced intentions to travel to a specific destination ( Hsu & Huang, 2012;Quintal et al., 2010;Shen et al., 2009;Sparks, 2007;Sparks & Pan, 2009). For instance, Sparks and Pan (2009) discovered that controlling resources like time and money was a notable predictor of visiting a target destination. Also, according to Sparks (2007), perceived control over time and expense was the most crucial predictor of why people wanted to go on a wine vacation. As a result, the following hypothesis is put forth: H4: The stronger an individual's perceived behavioural control over the potential of taking a vacation throughout the COVID-19 pandemic, the more likely he or she would be to do it.

Subjective Norm
According to Hsu and Huang (2012); Martin et al (2011);Quintal et al (2010); Sparks and Pan (2009); Lam and Hsu (2006) studies, there are significant associations exist between travel intentions and subjective norms based on social factors. Simply put, the participants in these research were more willing to visit the desired location if they believed that it would be an appealing or prudent action to do. Lam and Hsu (2006) verified the effect of reference groups on travel intentions. First and foremost, the researchers revealed that participants were more likely to heed the recommendations of family, friends, and relatives, and to a lesser extent, coworkers and travel agents. Eventually, Ziadat (2015) discovered that subjective standards had a direct and substantial effect on intentions to visit or revisit a destination. Consequently, the following hypothesis is formulated: H5: The higher an individual's vacation intention throughout the pandemic of COVID-19, the stronger the optimistic influence of the subjective norm on the vacation decision.

Willingness to Pay
Willingness to pay can be a method of determining the worth of items or services that lack a particular retail or trade worth. However, this compute is continually changing since it will change as individuals gather additional information (Nowacki, 2013). Previous studies showed that fear of flying boosted the WTP in order to improve safety (Koo et al., 2019). Preliminary research indicates that during the present COVID-19 pandemic, around one third of restaurant customers and 40 percent of hotel guests are willing to pay a higher premium for enhanced safety procedures . Although customers anticipate hospitality firms to implement more severe safety or cleanliness processes, some are ready to pay for the additional safety precautions . Therefore, the last hypothesis is proposed H6: The higher the intention to travel during the COVID-19 pandemic, the more individuals are prepared to spend for additional safety measures. Figure 1 depicts the suggested research model for this study, which is based on all hypotheses.

Population and Sample
The term "population" was first coined by Sekaran and Bougie (2013), who defined it as "the entire community of persons, situations, or interest concerns that the researcher wants to conclude about." This study's populations were the Malaysian 18 years old and above registered Facebook users in the travel group community of Cuti-Cuti Kaki Travel. This group was chosen because it has more than 143,000 followers (CCKT, 2021). Thus, it raised the number of potential respondents to answer the questionnaire and became the study's population.
In this study, the nonprobability sampling designs which fit into the broad categories of purposive sampling were used. Lavrakas (2013) mentioned that purposive sampling, often known as an expert or judgmental sample, is a form of nonprobability sampling. The basic objective of a purposive sample is to create a sample that may be trusted to reflect the population with some degree of confidence. This is also achieved by applying population expertise to pick a sample of elements that constitute a cross-section of the population in a non-random way. For this context, power analysis using G*Power 3.1 software was used to determine the appropriate sample size (Erdfelder et al., 2009). Based on this technique, the minimum sample required is 119. However, 181 samples have been collected considering the possibility of receiving unusable responses like too many missing values, incomplete responses, responses with significant outliers and others. This could prove that this study had to add another response to make it 181 out of 119 responses. These numbers can only be allowed to generalise the (sub) population from which the sample is drawn and not the entire study population (Andrade, 2021).

Research Instrument
The study data was obtained through a questionnaire. The questionnaire was designed in an e-survey (Google Form) that looks easy to answer and professionally designed. The  (2015); Zhang et al (2020) that comprises sixty-nine questions divided into two major sections; demographic information and related variables. Using a nominal scale, a series of questions were asked about gender, age, origin, education, employment/current activity, monthly household income, vacation frequency, vacation budget, preferred destinations and category of COVID-19. In the first section, demographic information was presented.
There were six parts of the questionnaire in the second section on the related variables. The first part was to measure the perceived risk (PR), the second part was to assess the attitude (ATT) and the third part was to measure the perceived behavioural control (PBC). Also, the fourth part was to assess the subjective norm (SN) and the fifth part was to measure the vacation intention (VI) and the sixth part was to assess the willingness to pay (WTP). Both English and Malay mediums were used to enhance understanding among respondents. Items were sent to an expert for Malay translation to ensure equality with English. Ensuring the instrument validity and reliability, pre-test and pilot study were carried out. The scale that used in the second section is a 5-point Likert scale.

Pilot Test
The surveys were distributed to actual respondents after pre-testing with experts. This was done to determine if the questions were clear, the wording was appropriate, and the amount of time it took to answer the questions was reasonable. This step was taken to help discover any potential flaws in the questions that could impact the study's outcomes in the future. The pilot test comprised 30 respondents who were willing to complete the questionnaires. For the reliability test, the data from the pilot test was entered and analysed using Statistical Package for the Social Science Software (SPSS) 21 st version. The pilot test verified that all constructs met the recommended Cronbach's Alpha value of above 0.6; PR (0.839), ATT (0.951), PBC (0.922), SN (0.932), VI (0.950) and WTP (0.969). According to Hair et al (2019), a Cronbach Alpha value of more than 0.6 was deemed sufficient for reliability.

Data Collection
The data were collected from June until July 2021 within this current pandemic situation. The researcher distributed the questionnaire through the e-survey method using Google Form to the potential respondents at Cuti-Cuti Kaki Travel's Facebook travel group community (CCKT, 2021). Initially, the purpose was to collect data from a single Facebook travel group community of Cuti-Cuti Kaki Travel (CCKT, 2021). However, the response rate was extremely low due to low Facebook group activity. Therefore, the questionnaire was distributed to all prospective Facebook travel group communities such as Cuti The researcher contacts the group administrators through Facebook messenger first for permission approval. Once obtained approval, the questionnaire was distributed to a group member using the Google Form link. The choice of e-survey is made due to the government's limitation of movement restriction (Majlis Keselamatan Negara [MKN], 2021). A brief introduction was given to the respondents regarding the research and the questionnaire. The respondents needed to answer the e-survey through a questionnaire prepared in Google Form and it was anticipated that it would take less than 10 minutes to complete the e-survey. There was a total of 181 valid responses. Some questionnaires were eliminated because they were incomplete or erroneously filled out, resulting in this final database. The data collection started in June 2021 and it took a month to complete the process.

Data Analysis
Partial Least Square -Structural Equation Modelling (PLS-SEM) with SmartPLS 3.3 software was used to examine the influence of TPB predictors and perceived risk on vacation intention and willingness to pay among domestic vacationers in Malaysia during the COVID-19 pandemic. The PLS-SEM approach was chosen for its adaptability in predicting causal relationships with small sample data and the fact that it requires almost no assumptions about data distributions. The PLS algorithm was used to examine the reliability and validity of the survey items and constructs. Items with an outer loading value of less than 0.70 are removed from the construct. After the measurement model has been examined, the structural model is evaluated using bootstrapping with 5,000 samples and the blindfolding technique. The path model keeps structural paths that have large loadings. The answer to the research objectives comes from significant paths.

Results
The Measurement Model's Procedure, Validity and Reliability Table 1 shows the respondent's demographic information. The PLS-SEM approach was used to test the model and hypotheses stated. Internal consistency, convergent validity, and discriminant validity (Table 3) of the measurement model were confirmed, and the loadings of the each indicators were analysed to determine their realibility (Table 2). Due to the unavoidable multicollinearity issue, a number of items in various constructs had to be removed from the measurement model, resulting in the reduction of PR to six items and PBC to eight indicators.
Except for five items pertaining to PR, ATT and SN, the majority of the outer loadings surpass the suggested cut-off value of 0.707 (Carmines & Zeller, 1979). However, since they are unrelated constructs that do not fall below the 0.4 cut-offs set by Hair et al (2019) for eliminating an indicator, they are kept in the model since they can assist in extracting meaningful information from the indicator in order to create a more robust score of latent variable and validate discriminant validity based on the remainder of the measurement indicators (Table 3). As recommended by Nunnally and Bernstein (1994), the Composite reliability surpasses the 0.7 thresholds as an appropriate degree of reliability in terms of internal consistency and convergent validity ( Table 3). All of the constructs meet the Fornell and Larcker (1981) average of variance extracted (AVE) criterion by surpassing the cut-off of 0.5 in terms of convergent validity; specifically, elucidate every construct explicates at least 50 percent of the variance of the allotted indicators. The square root of the variance shared by the construct and its indicators (AVE), represented by the values on the main diagonal (in bold), is significantly more prominent than the associations between each construct and any other construct (the remainder of the matrix), demonstrating discriminant validity. Figure 2 displays the estimate and validation of the structural model. The parameters were estimated using a bootstrap approach with 5,000 samples after verifying the dearth of multicollinearity (VIF < 5 for all indicators); they were notable in all situations (p < 0.05), with the exception of the path between VI and WTP ( Table 4). The standardised root mean square residual (SRMR) is 0.075, which indicates the model's goodness of fit because it does not exceed the cut-off of 0.10 (Henseler et al., 2009).  Except for the correlation between VI and WTP (R 2 = 0.478), the R 2 values of the endogenous variables exceed the threshold value of 0.1 (Falk & Miller, 1992). The direct correlation of VI with ATT (R 2 = 0.026) and with PBC (R 2 = 0.033) exhibit weak effects since they do not exceed the 0.33 threshold value (Chin, 1998). The cumulative effect of ATT and PBC on VI might be deemed weak (R 2 = 0.059).

Hypothesis Testing
Both ATT toward VI (β = 0.176, p > 0.019) and PBC (β = 0.197, p > 0.006) are positively influenced by PR. Since both constructs are positive and statistically significant, hypotheses 1 and 2 can be approved.
Assessment of the direct effect of the three explanatory variables of VI, in terms of the TPB, have a major and direct positive effect on it. The construct associated with ATT has a value that is only significantly lower than PBC (β = 0.255, p > 0.006), indicating that a positive attitude promotes VI. Meanwhile, PBC (β = 0.317, p < 0.000) had the biggest quantitative effect, implying that if a person believes he or she has control over the elements connected with vacationing, he or she will be more likely to vacation. The SN (β = 0.261, p < 0.004) is a positive and statistically significant construct. The opinion of the referents of an individual on the notion of vacationing has a substantial influence on VI. In any event, the model supports hypotheses 3, 4 and 5.
Finally, hypothesis 6 is accepted because the construct related to the effects of VI on people's WTP for increased safety features throughout the journey is statistically notable (β = 0.503, p < 0.000). To put it another way, having a stronger vacation intention indicates that the person would be prepared to spend more than the regular price to be certain of enhanced safety precautions as a result of COVID-19. Therefore, hypotheses 1, 2, 3, 4, 5, and 6 were supported and accepted. This result would be investigated further in terms of its significance for the tourism industry.

Discussions and Recommendations
In the context of theoretical implications, the proposed model integrates the perceived risk of vacationing in the midst of the COVID-19 pandemic, which has a substantial impact on both attitudes and perceived behavioural control (PBC). This is consistent with previous claims that risk perception influences vacation intentions via these dimensions. It should be noted, however, that other research in pandemic settings did not find a correlation between perception and vacation intention, suggesting that individuals planned to implement adaptive behaviours to reduce the risk of infection.
Therefore, future research should investigate further the role of perceived peril and risk in pandemic vacation intentions. Understanding the various factors that contribute to risk perception and how individuals adapt their behaviour can provide tourism industry stakeholders with valuable insights.
In addition, the study's TPB model indicates that PBC has the greatest influence on vacation intention, followed by attitudes. In a limited capacity, subjective norms also play a role. This finding is consistent with prior research highlighting the impact of social norms and personal beliefs on vacation intentions.
Thus, tourism agents and policymakers should prioritise strategies that increase the sense of control individuals have over their vacation experiences. Potential vacationers can be attracted by emphasising empowerment, emphasising the positive aspects of the vacation, and addressing risk perceptions. Additionally, efforts should be made to convey the safety precautions, hygiene protocols, and quality standards implemented by tourism service providers.
Moreover, the notion of willingness-to-pay (WTP) for stricter safety precautions in vacation-related activities during the pandemic is introduced in this study. More respondents who plan to take a trip reported being willing to pay more for greater security.
Therefore, providers of recent tourism-related services should consider incorporating and communicating enhanced safety measures. Event though in recent years, the active COVID-19 cases are not as much as the period where the virus begin to strikes the world before, customers can be reassured and their trust can be established by emphasising the investment in risk reduction measures and including the additional costs in the price of the product or service.
As for practical implications point of view, participants in the tourism industry should concentrate their marketing efforts on the variables that have the greatest influence on vacation intention: PBC, subjective norms, and attitudes. While attitudes have less of an impact on people's desire to travel, the opinions of friends and family play a significant role.
To support the above idea, agents in the tourism industry should utilise social networks and referral systems to influence vacation plans. As of 16 th of June 2023, the data of 18,147 active COVID-19 cases provided by Ministry of Health (Ministry of Health, 2023), potential vacationers should be provided with pertinent information about their capacity, opportunity, and desire to travel despite the pandemic as part of promotional campaigns. Utilising social media platforms and collaborating with travel operators can help impart a sense of safety and quality in the tourism environment.
Besides, accreditation for tourism quality, brand reputation, and information about hygiene, safety, and sanitation initiatives are essential for lowering psychological barriers to risk and enhancing the perception of control among individuals. Thus, tourism service providers should invest in relevant accreditations and certifications that demonstrate their commitment to safety and cleanliness. Communicating these measures to prospective vacationers through a variety of marketing channels can alleviate their concerns and enhance their sense of control.

Future Lines of Study
The study acknowledges the small sample size obtained from a solitary Facebook travel group community. Multiple Facebook groups were required to distribute questionnaires due to low response rates. This may have an effect on the generalizability of the results.
Therefore, future studies should strive for larger sample sizes to increase the representativeness of the findings and the statistical power of the investigation.
In addition, research conducted in various countries or phases of the pandemic can provide valuable insights into how cultural contexts and varying pandemic situations affect travel intentions and risk perceptions.
Thus, to comprehend the nuances of vacation intentions during pandemics, researchers should consider conducting comparative studies across multiple nations. Examining the evolution of risk perception throughout the various phases of a pandemic especially in the recent years can also provide a deeper understanding of individuals' decision-making processes.
Lastly, according to the findings, future research should look into other theoretical frameworks and approaches to acquire a better understanding of vacation intentions during pandemics. For instance, other behaviour theories or frameworks that could complement or enhance the TPB model should be investigated by researchers. Using a variety of research strategies, such as qualitative approaches or longitudinal studies, can provide new views and insights about vacation intentions.

Conclusion
As conclusion, this study contributes to the existing literature on comprehending vacation intentions during the COVID-19 pandemic with the proposed study that integrates perceived risk, attitudes, subjective norms, and perceived behavioural control (PBC). Perceptions of risk influence vacation intentions via attitudes and PBC, although some studies suggest that adaptive behaviours can reduce infection risk. Thus, future research could as well investigate further the role of perceived risk and danger in pandemic vacation intentions, providing tourism industry stakeholders with valuable insights.
The practical implications of this study emphasises the significance of PBC, attitudes, and subjective norms in determining vacation intentions, with PBC having the greatest impact. Tourism agents and policymakers should prioritise strategies that enhance the sense of control of individuals, highlight the positive aspects of vacations, address risk perceptions, and communicate safety measures. The willingness to pay for stricter safety precautions is also introduced, highlighting the significance of implementing and communicating enhanced safety measures. Marketing efforts should focus on PBC, subjective norms, and attitudes, utilizing social networks, referral systems, and pertinent information about travel capacity and opportunities. Accreditation for quality, brand reputation, and hygiene measures can reduce psychological barriers and increase the sense of control in individuals. Therefore, future research could aim for larger sample sizes, investigate cross-cultural contexts and pandemic phases, and consider alternative theoretical frameworks and research methods in order to acquire a better understanding of vacation intentions during pandemics.