Personality Traits That Affect Travel Intentions during Pandemic COVID-19: The Case Study of Serbia

The pandemic caused by COVID-19 has enormous effects on peoples’ lives and on the global economy. The outbreak and effects of COVID-19 have resulted in fears, concerns, and anxiety among people all around the world. The aim of this study is to investigate the relationship between the fear of COVID-19, travel anxiety, and travel intention among residents in Serbia. In addition to the above, the aim of the study was to examine the impact of personality traits on the fear of COVID-19, travel anxiety, and travel intentions. To enhance the understanding of the relationship between the mentioned variables, this study proposes a research model based on the use of measurement scales from the existing studies to develop the questionnaire. The standard paper-and-pen survey was used for data collection from the respondents from Serbia. The total sample include 987 respondents. It was determined that Travel intention was negatively influenced by Neuroticism and Conscientiousness, but positively by Extroversion and Openness. The supposed Agreeableness did not have a significant effect on Travel intention. This study contributes to understanding the respondents’ behavior during pandemic of COVID-19, and their responses to the crisis according to their personality traits might have practical implications in the travel sector.


Introduction
The pandemic caused by COVID-19 has been spreading throughout China and worldwide; the pandemic danger has not only threatened lives due to infection, but has also brought about enormous psychological tension [1]. Besides adverse effects on peoples' physical and mental health, the outbreak of COVID-19 has proven to be an economic problem [2]. All the sectors, without exception, suffer massive damage, but the extent of the damage varies among sectors and countries [3]. Compared to SARS which spread to more than 20 nations in 2003, more people are infected by the new coronavirus [4][5][6]. With the aim to protect the population and prevent the spread of the pandemic, drastic measures were taken by every country in the world. All the countries closed schools, borders, and stopped production and trade. These changes had effects on all the areas of economy, including the tourism industry [5,7,8]. There are too many ways in which pandemics can affect tourism and hospitality industry. Primarily, restrictions that are related to travel have a negative impact on economy in general, not only on the tourism and hospitality industry. Travel restrictions which have been implemented to limit the spread of the new coronavirus and to protect residents' health [9,10] made it difficult for individuals to move. As stated by

Instruments
The questionnaire used in this study was composed from several scales used in previous research which has demonstrated validity and significance of the assessment scales. All the items applied in this study were back-translated in Serbian by the authors and the accuracy of translation was verified. For assessing the level of agreement, respondents marked answers from 1-strongly disagree to 5-strongly agree (5-point Likert scale).
Sociodemographic variables: Questions referring to sociodemographic characteristics were placed in the first part of the questionnaire. Examinees were asked to provide answers about their gender, age, marital status, education level, and monthly incomes. The last question referred to their traveling during the pandemic.
Fear of COVID-19 (F) was measured by a 7-item scale adopted from Ahorsu et al. [30] and Luo and Lam [45]. The scale included items such as "My heart races or palpitates when I think about getting COVID-19". The Fear of COVID-19 total score also suggests that this is a good psychometric tool that can be compared by gender and all age groups and can be used to assess and mitigate the fear of pandemics among individuals [30].
Travel Anxiety (TA): Six-item model adopted from Wachyuni and Kusumaningrum [94] assessed the respondents' concerns related to traveling after proclaiming the pandemic of COVID-19 to be over. An example of the item is "I feel an irregular heartbeat when I think of going on a tour during the pandemic".
Travel Intention (TI): Three-item scale was used for measuring travel intention of the respondents. This scale was adopted from Luo and Lam's study [36]. In the original study, travel intention referred to traveling to travel bubble destination. For the purpose of this study, the items were modified. Considering the current situation of spreading COVID-19 and the number of new cases in the countries of the region, the authors decided to examine travel intention to countries with the similar situation as in Serbia. Based on that, the example of the item is "I would like to travel to countries with a similar epidemiological situation in the future".
BFI-Five major types of personality traits have been built into commonly recognized and useful assessments for investigating, classifying and predicting behavior. Many of these scales are based on the five-factor model, which is often known as the Big-Five [95][96][97]. The BFI self-report (BFPTSQ) [98] measure involves 44 items and evaluates Agreeableness (A), Extroversion (E), Conscientiousness (C), Openness (O), and Neuroticism (N).

Procedure
The data were collected through a standard paper-and-pen survey, utilizing a convenient method for sampling. The survey was conducted from December 2020 until July 2021, and the participation was anonymous and voluntary. Adult citizens of the Republic of Serbia (18 and older) were included in the research. Data collecting was conducted in larger city centers of Serbia (Belgrade, Novi Sad, Niš, Kragujevac, Subotica, Zaječar,Čačak, Leskovac, Novi Pazar, and Vranje). So as to obtain a relevant sample that included the citizens from all the parts of the country, the city centers from all the regional entities of the Republic of Serbia, excluding Kosovo and Metohija, were chosen [99]. For conducting the survey, 20 students from the Faculty of Sciences, University of Novi Sad, were engaged. There was a total of 1500 survey questionnaires distributed (150 in each city). The questionnaires were distributed in shopping centers and hospitality facilities. Of the total number, 987 of them were properly filled out (the response rate of 65.8%). The average time to fill out the questionnaire was 15 min. These data were processed by R and RStudio (lavaan, semPlot, and semTools packages) which was used for the CFA and SEM analyses.

Study Sample
The sample consisted of 987 participants. There was a higher number of women in the sample (55.9%), while the average age of the participants was 35.78 (age range . Most of the respondents have an income in the range 40,001-50,000 RSD (33.2%), hold a BSc degree (47.5%), and are married (45.8%). When it comes to traveling during COVID-19, 64.8% said that they did not travel during the pandemic (Table 1).

Measurement Model Validity-Confirmatory Factorial Analysis
The latent factors measurement model was estimated to check for innate construct validity and reliability using Confirmatory factorial analysis (CFA). Table 2 shows the results of the measurement model estimation. According to the results, all fit indices (TLI = 0.987; CFI = 0.986; RMSEA = 0.075; SRMR = 0.062) revealed a satisfactory fit.  The reliability of the scale was assessed through average variance extracted (AVE) indices and composite reliability (CR) ( Table 2). The convergent validity of each dimension was examined by calculating the score of the average variance extracted (AVE). A substantial convergent validity is achieved when all item-to-factor loadings are significant and the AVE score is higher than 0.50 within each dimension [100]. The composite reliability values for the latent factors exhibited satisfactory levels exceeding the recommended minimum value of 0.7 [101,102]. The results indicate that all the dimensions had AVE higher than 0.50 and CR higher than 0.70 (Table 2), which indicates good convergent validity. The results indicate that the alpha coefficients of the nine factors, with radius from 0.854 to 0.968, which demonstrates that the scales of the questionnaire, had considerable reliability [103] ( Table 2).
The discriminant validity was assessed using Fornell and Larcker [100] by comparing the square root of each AVE in the diagonal with the correlation coefficients for each construct in the relevant rows and columns ( Table 3). The AVE values' square roots have a greater value than the correlations with other latent constructs. Thus, the results confirm that all the dimensions have sufficient discriminant validity [81], which is shown in Table 3.

Results of the Path Model
The variety of forms in the proposed model were assessed and the overall fit of the model was considered satisfactory for the observed sample (CFI = 0.951; TLI = 0.946; RMSEA = 0.051; SRMR = 0.079; df = 18, p < 0.000). Fit indices were acceptable for addressing the hypothesized interrelations between each latent factor. Table 4 shows the outcomes of the hypothesized interactions in the proposed model.

Results of the Path Model
The variety of forms in the proposed model were assessed and the overall fit of the model was considered satisfactory for the observed sample (CFI = 0.951; TLI = 0.946; RMSEA = 0.051; SRMR = 0.079; df = 18, p < 0.000). Fit indices were acceptable for addressing the hypothesized interrelations between each latent factor. Table 4 shows the outcomes of the hypothesized interactions in the proposed model.

Results of the Path Model
The variety of forms in the proposed model were assessed and the overall fit of the model was considered satisfactory for the observed sample (CFI = 0.951; TLI = 0.946; RMSEA = 0.051; SRMR = 0.079; df = 18, p < 0.000). Fit indices were acceptable for addressing the hypothesized interrelations between each latent factor. Table 4 shows the outcomes of the hypothesized interactions in the proposed model.

Results of the Path Model
The variety of forms in the proposed model were assessed and the overall fit of the model was considered satisfactory for the observed sample (CFI = 0.951; TLI = 0.946; RMSEA = 0.051; SRMR = 0.079; df = 18, p < 0.000). Fit indices were acceptable for addressing the hypothesized interrelations between each latent factor. Table 4 shows the outcomes of the hypothesized interactions in the proposed model.

Results of the Path Model
The variety of forms in the proposed model were assessed and the overall fit of the model was considered satisfactory for the observed sample (CFI = 0.951; TLI = 0.946; RMSEA = 0.051; SRMR = 0.079; df = 18, p < 0.000). Fit indices were acceptable for addressing the hypothesized interrelations between each latent factor. Table 4 shows the outcomes of the hypothesized interactions in the proposed model.

Results of the Path Model
The variety of forms in the proposed model were assessed and the overall fit of the model was considered satisfactory for the observed sample (CFI = 0.951; TLI = 0.946; RMSEA = 0.051; SRMR = 0.079; df = 18, p < 0.000). Fit indices were acceptable for addressing the hypothesized interrelations between each latent factor. Table 4 shows the outcomes of the hypothesized interactions in the proposed model.

Results of the Path Model
The variety of forms in the proposed model were assessed and the overall fit of the model was considered satisfactory for the observed sample (CFI = 0.951; TLI = 0.946; RMSEA = 0.051; SRMR = 0.079; df = 18, p < 0.000). Fit indices were acceptable for addressing the hypothesized interrelations between each latent factor. Table 4 shows the outcomes of the hypothesized interactions in the proposed model.

Results of the Path Model
The variety of forms in the proposed model were assessed and the overall fit of the model was considered satisfactory for the observed sample (CFI = 0.951; TLI = 0.946; RMSEA = 0.051; SRMR = 0.079; df = 18, p < 0.000). Fit indices were acceptable for addressing the hypothesized interrelations between each latent factor. Table 4 shows the outcomes of the hypothesized interactions in the proposed model.

Results of the Path Model
The variety of forms in the proposed model were assessed and the overall fit of the model was considered satisfactory for the observed sample (CFI = 0.951; TLI = 0.946; RMSEA = 0.051; SRMR = 0.079; df = 18, p < 0.000). Fit indices were acceptable for addressing the hypothesized interrelations between each latent factor. Table 4 shows the outcomes of the hypothesized interactions in the proposed model.

Results of the Path Model
The variety of forms in the proposed model were assessed and the overall fit of the model was considered satisfactory for the observed sample (CFI = 0.951; TLI = 0.946; RMSEA = 0.051; SRMR = 0.079; df = 18, p < 0.000). Fit indices were acceptable for addressing the hypothesized interrelations between each latent factor. Table 4 shows the outcomes of the hypothesized interactions in the proposed model. Fear of COVID-19 on Travel anxiety (H1) indicated significant and positive impact (β = 0.693, p < 0.000). Fear of COVID-19 (β = 0.011, p = 0.759) and Travel anxiety (β = −0.006, p = 0.866) were not found to negatively affect the Travel intention (H2 and H3). Extroversion

Results of the Path Model
The variety of forms in the proposed model were assessed and the overall fit of the model was considered satisfactory for the observed sample (CFI = 0.951; TLI = 0.946; RMSEA = 0.051; SRMR = 0.079; df = 18, p < 0.000). Fit indices were acceptable for addressing the hypothesized interrelations between each latent factor. Table 4 shows the outcomes of the hypothesized interactions in the proposed model.

Results of the Path Model
The variety of forms in the proposed model were assessed and the overall fit of the model was considered satisfactory for the observed sample (CFI = 0.951; TLI = 0.946; RMSEA = 0.051; SRMR = 0.079; df = 18, p < 0.000). Fit indices were acceptable for addressing the hypothesized interrelations between each latent factor. Table 4 shows the outcomes of the hypothesized interactions in the proposed model.

Results of the Path Model
The variety of forms in the proposed model were assessed and the overall fit of the model was considered satisfactory for the observed sample (CFI = 0.951; TLI = 0.946; RMSEA = 0.051; SRMR = 0.079; df = 18, p < 0.000). Fit indices were acceptable for addressing the hypothesized interrelations between each latent factor. Table 4 shows the outcomes of the hypothesized interactions in the proposed model. Fear of COVID-19 on Travel anxiety (H1) indicated significant and positive impact (β = 0.693, p < 0.000). Fear of COVID-19 (β = 0.011, p = 0.759) and Travel anxiety (β = −0.006, p = 0.866) were not found to negatively affect the Travel intention (H2 and H3). Extroversion

Results of the Path Model
The variety of forms in the proposed model were assessed and the overall fit of the model was considered satisfactory for the observed sample (CFI = 0.951; TLI = 0.946; RMSEA = 0.051; SRMR = 0.079; df = 18, p < 0.000). Fit indices were acceptable for addressing the hypothesized interrelations between each latent factor. Table 4 shows the outcomes of the hypothesized interactions in the proposed model.

Discussion
The study results provided a better insight into people's responses to the pandemic depending on personality traits. Individuals' behavior during the pandemic might be predicted if we have information about their personality characteristics.
Although fear and anxiety are considered as two distinct feelings [104], many studies found a positive relationship between them [30,45,46]. Based on that, it was assumed that increase in Fear of COVID-19 will increase Travel anxiety (H1). The results of the study confirm the authors' expectations and gave support for H1. Although it was expected to find a negative connection between Fear of COVID-19 and Travel intention, the results show no notable connection between these two variables. Even though these results do not provide support for H2, they confirm the results of Luo and Lam's [45] study. Therefore, there is no indication that fear related to COVID-19 can decrease people's intention to travel. Perhaps it can be explained through the fact that COVID-19 is present worldwide and the risk of getting infected is possible in their countries as well as in other tourist destinations. Furthermore, based on the existing literature, it was expected to find a negative correlation between Travel anxiety and Travel intention [45,105], but the effects of Travel anxiety on Travel intention are not significant, which rejected H3. Due to the increase in the number of people infected by COVID-19, on 15 March 2020, a state of emergency was declared in the territory of the Republic of Serbia, which was revoked on 6 May 2020 [106]. After the lifting of the restrictive measures and lockdown, the number of visitors in tourist destinations in Serbia increased, and the number of domestic tourists rose by 26.66% in the first six months, compared to the same period in the previous year. The fastest recovery after the declaration of the state of emergency was recorded in spa and mountain tourist centers in Serbia, which means that the prevailing motive was relaxation, leisure, sports, and recreation [107,108]. The research carried out by Abdullah [109] confirmed that the tourists decided to travel stirred by "revolt and boredom" during the pandemic because they were excluded from social activities during the period of isolation. The aforementioned implies that although there was fear of infection with COVID-19, it did not affect the intention to travel among the examinees in the observed sample.
Extroversion. Considering the fact that the characteristics of Extroversion can help in recuperating from negative effects of stressful situations [110], it was expected to find a negative relationship between Extroversion and Fear of COVID-19. The obtained results confirm H4. Since Extroversion was found to be negatively correlated with COVID-19 anxiety syndrome [89], it was supposed that Extroversion negatively affects Travel anxiety (H9). The results show that there is no significant influence of Extroversion on Travel anxiety, which rejected H9. The alternative explanation could be reflected in the fact that socializing with others is how extroverts recharge their batteries. Among extroverts, months of quarantine and various protective measures affected the need for socialization, including the need to travel. Insignificant results can be explained through the fact that extroverts rushed to travel after the lifting of the measures, without developing a state of anxiety from traveling. If we consider the abovementioned facts, it was assumed that persons with extroversion traits have higher intentions to travel (H14). The results show significant positive effects of Extroversion on Travel intentions, which is in accordance with previous research [84], and confirms H14.
Agreeableness. In previous studies, a negative association has been found between Agreeableness and social distancing [111], which indicated their characteristics of generosity, helpfulness, and empathy [112]. People who score high in Agreeableness like to volunteer and to make social contacts, which increases their likelihood to be infected. Based on this, it was assumed that people who had agreeableness characteristics will express lower Fear of COVID-19. The results show negative correlation between these two variables and provide support for H5. In continuation of this, it was expected that those who score high in Agreeableness would express lower Travel anxiety and higher Travel intention. By testing mutual influences, it was obtained that there were no significant effects of Agreeableness on Travel anxiety (H10) and Travel intention (H15). Since no presumed connections were obtained, the authors believe that an alternative explanation could be found in the fact that people with more agreeable personality are more likely to function well in the restricted settings during lockdowns and social distance limitations [113], which may justify the absence of a significant relationship between the variables tested.
Neuroticism. Since Neuroticism is correlated with most of the emotional disorders [114,115], it was expected that people with higher scores of Neuroticism will express higher Fear of COVID-19 (H7), as well as higher Travel anxiety (H12). Our results indicate strong effects of Neuroticism on Fear of COVID-19, which confirms H7. These results confirm the earlier findings obtained in similar studies related to coronavirus epidemics [35,85,116]. In addition to this, Neuroticism was positively correlated with Travel anxiety (H12), which confirmed the results from previous studies and provided support for H12. The results show negative correlation between Neuroticism and Travel intention which provided support for H17. Even in the case of different results, it would not be surprising and difficult to explain that when coping with a pandemic disease, persons who are prone to neuroticism may exhibit inconsistency in their behavior, which may be related to their level of neuroticism. Persons who express higher levels of fear and anxiety, for example, prefer to use denial as a psychological defense strategy to alleviate their fears; this implies that they participate in risky activities to feel psychologically safe [117][118][119]. This practically means that they would dare to travel even though it is risky.
Conscientiousness. Like the people with neuroticism, those who have conscientiousness traits can be related with conforming to social standards and engaging in social distancing [111] and handwashing [120] during the pandemic of COVID-19. Hence it was expected that Conscientiousness has positive correlation with Fear of COVID-19 (H6) and Travel anxiety (H11). The results show that people who score high in Conscientiousness express higher Fear of COVID-19 and Travel anxiety, which confirmed H6 and H11. Higher Conscientiousness is associated with higher likelihood of taking preventing measures [111,120] which might negatively affect Travel intention, which was confirmed in this study (H16). These results are not surprising due to the fact that people who score higher in Conscientiousness look ahead and consider how their actions influence other people. They do not make rash or thoughtless decisions, which is also applicable in the case of behavior and travel decisions during coronavirus.
Openness. In previous studies, the negative relationship between Openness, Agreeableness, Extroversion and anxiety, negative feelings, and depression [121,122] was deter-mined. Based on that, it was expected that Openness negatively affects Fear of COVID-19 (H8) and Travel anxiety (H13). The results reject both hypotheses because there is no notable relationship. The possible explanation which the authors could offer is the fact that the coronavirus has been present in people's lives for a long time, and that they have become accustomed to living with it. The assumption is that the initial fear of infection gradually decreased over time, which may vindicate truancy of a significant impact of Agreeableness on Fear of COVID-19 and Travel anxiety. According to Götz et al. [123], people who are open are less conformist and tend to strive for experiences that are unconventional [124], hence it was expected that Openness would positively affects Travel intention (H18). Since analysis showed positive impact of Openness on Travel intention, the H18 is supported.

Conclusions
One of the most significant and unfortunate occurrences of contemporary times was the outbreak of COVID-19. This led to a serious economic crisis, and tourism and hospitality industry are the most impacted. Pandemic future development is still unknown which makes it difficult to assess the impact on tourism and hospitality. In the conditions of global pandemic, tourism and hospitality are under the influence of the factors which shape users' behavior, such as: perceived health risk, preventive measures prescribed by the government, their income, or the fear of getting infected. The mentioned factors may be perceived in different ways depending on the consumers' personality traits. The results obtained in this research proved that some of the personality traits can have a significant impact not only on travel intentions, but also on the fear of the global pandemic, as well as on the occurrence of travel anxiety. In this study, 18 hypotheses were tested, eleven of which were confirmed. The greatest negative impact on the travel intentions of the respondents from the urban centers of Serbia was the one of the Conscientiousness, while the greatest positive impact was confirmed for Extroversion. In order to preserve the economic stability of tourism and hospitality companies, it is necessary to respect all the suggested measures which will not only reduce the risk of infection transfer, but they will also contribute to the creation of the atmosphere in which tourists feel safe, which will further mitigate their anxiety and fear.
Finally, this study is important because there were no instruments available to measure Fear of COVID-19 in Serbia. Although this scale has been validated in many languages e.g., [46,[125][126][127][128][129], this study established Fear of COVID-19 scale (FCV-19S) as a confident and valid tool for evaluating fear associated with COVID-19 on the Serbian sample, which reflects the main contribution of the conducted study.

Limitations and Future Research
There are several limitations that should be considered. First of all, the questionnaire used in this research was based on self-report measures, which can result in more subjective perception of the measured variables. Secondly, the sampling was carried out based on the principle of a random appropriate sample, so the obtained results cannot be generalized. Moreover, the sample included only the respondents in urban centers, which points to the need to expand the research to the rural areas of Serbia as well. Additionally, the questionnaire should contain additional information about the respondents that relate to their health state and the health state of their family members in order to investigate whether the fear of COVID-19 has a significant connection with vulnerable groups in the population (e.g., patients with chronic diseases, pregnant women, elderly people). In previous studies, the impact of demographic factors was identified. For example, women expressed higher levels of emotional distress and negative thoughts related to COVID-19 compared to men [122,130,131]. Additionally, other reports indicated that parents are more likely to become more concerned [132], because they tend to worry not just about the health of their families and their own health, but about the health of their family and children [122,131]. Hence, due to the established importance of sociodemographic factors in coping with the pandemic caused by COVID-19, the recommendation for future research is to include these variables in the analyses in order to better understand the reactions to COVID-19. Finally, the further course of pandemic is uncertain. The epidemiological image in the world is changing every day, as well as the preventive measures related to the process of traveling and staying in hospitality facilities. Due to this, it is necessary to perform continuous research of tourists' attitudes so that the sector of tourism and hospitality should have a timely response to the challenges placed by the global pandemic.