Travel risk perceptions and behavior in the course of the COVID-19 pandemic 2021: a cluster analysis

ABSTRACT

This paper discusses the impact of the COVID-19 pandemic on tourism destinations in Uruguay by exploring travel behavior. Through a survey carried out between DEC2020 and JAN2021, the study reveals that despite the pandemic, the desire for holiday tourism remained. The empirical results show that, despite the fact that the vast majority have a high perception of the risk of COVID-19, there are differentiated behaviors regarding the decisions to travel and the health measures to be taken. Depending on their socio-demographic characteristics, respondents behave differently when faced with risk, and the study identifies groups more willing to travel within the country and abroad in times of a pandemic, compared to others that are more moderate and cautious about risking travel. Due to the absence of visitor databases, this sample is based on the convenience sampling method and is composed of individuals engaged or interested in Uruguay as a destination. Although this research does not allow for extensions or extrapolations to a wider set of respondents, it does provide relevant demand-side information to tourism service providers in Uruguay, such as how to shape services and products to respond to the level of risk shown by this study. Moreover, the special period in which the survey was administered, coinciding with the implementation of measures by the Uruguayan government to manage the pandemic, could on the one hand be considered a bias in the analysis, but on the other hand it sheds light on the influence of these measures on demand behavior.

KEYWORDS:

• Cluster analysis
• sun & beach tourism
• risk
• COVID-19
• destination management
• Uruguay

Introduction

COVID-19 was first detected on 31 December 2019 in Wuhan (Hubei Province), China. On 30 January 2020, the World Health Organization (WHO) announced the outbreak of the respiratory disease Coronavirus 2019 (COVID-19) caused by the ‘SARS-CoV-2’ virus. By March 2020, COVID-19 had affected more than 70 destinations worldwide. On March 18, there were a reported 191,127 COVID-19 cases throughout the world with 74,760 cases in Europe alone (WHO, 2020). Tourism has been one of the sectors most affected by the COVID-19 outbreak. The United Nations World Tourism Organization (UNWTO) predicted a decrease in tourist arrivals in 2020 compared to 2019 of 60–80%, based on a decrease in arrivals of 57% for the month of March. UNWTO notes that the impact of COVID-19 on tourism will be much higher than the recession following the Great Financial Crisis of 2008 when international tourist arrivals dropped by 4%, and the earlier SARS crisis, which affected the industry only by 0.4% internationally (UNWTO, 2020). Countries around the world have taken drastic measures to block areas, cities and even close international borders for extended periods, negatively affecting the global tourism industry, particularly the travel and accommodation sectors. Epidemic crises can cause major changes in demand for certain destinations, as travelers may consciously decide not to expose themselves to such catastrophes (Fotiadis et al., 2021; Moosa & Khatatbeh, 2021; Seraphin, 2021).

The increased perception of travel risk during COVID-19 is consistent with previous research on infectious disease outbreaks (Brida et al., 2021; Cahyanto et al., 2016; Lee et al., 2012; Leggat et al., 2010; Neuburger & Egger, 2021; Pine & McKercher, 2004). As mentioned above, the intention to avoid or cancel trips during a pandemic such as COVID-19 is highly related to the perception of risk to travel in general and especially to destinations with reported cases (Cahyanto et al., 2016) higher perceived susceptibility infection with COVID-19 while traveling (de Zwart et al., 2007) and self-efficacy leading to actions to mitigate any risk and avoid travel (Liao et al., 2010). This is consistent with the results of other studies, which show that the perception of health risks is negatively related to the perception of safety of a destination and can influence the travel intentions of travelers (Reisinger & Mavondo, 2006 Yoo et al., 2022;). It is possible that the spread of COVID-19 is also related to geography (Gozgor et al., 2022, 2021), with small countries and islands having some advantages over the spread of COVID-19; however, these countries may be more vulnerable to uncertainty shocks related to COVID-19.

To implement future actions for the recovery of the tourism sector in the region, information on the perception of risk will help to implement resilience policies. Gaining the trust of consumers will be essential so as to reactivate the debilitated tourism industry (Sharma et al., 2021). The revival of the tourism industry will depend on increasing confidence in travel and reducing the perception of the risk involved (Assaf & Scuderi, 2020).

Gozgor et al. (2021) draw a parallel between pandemic risks and geopolitical risks, positing that the current uncertainty related to COVID-19 may change the effects of geopolitical risks on tourism supply in various ways including restrictions on movement of people and goods between countries, which affects the demand for tourism and, ultimately, the tourist offer.

Neuburger and Egger (2021) used cluster analysis to conduct a study on the perception of travel risk and the travel behavior of travelers in the face of COVID 19 in the DACH region (Germany, Switzerland and Austria) two weeks before and immediately after the COVID-19 pandemic was declared (4 March and 19 March 2020, respectively). Using the questions on risk perception they concluded that COVID-19 risk perception, travel risk perception, and willingness to change or cancel travel plans increased significantly over the two-week period.

The purpose of this study is to identify clusters of tourists regarding travel during the COVID-19 outbreak in 2020 based on their perception of COVID-19, travel risk perception and travel behavior, and how it influences travel behavior in Uruguay. More specifically, the objective is to examine the change in the perception of travel risk and the segments of the traveler profile of the perceived risk of COVID-19 during the pandemic outbreak at the beginning of the summer season in Uruguay, that is the peak of demand for tourist services. This study provides a novel perspective on the development of risk perception and travel behavior during a health crisis by identifying the profiles of potential tourists during a pandemic outbreak and seeking to deepen the understanding of the relationship between risk perception and travel behavior.

Risk perception can be explained as the subjective assessment of the risk of a threatening situation based on its characteristics and severity (Moreira, 2008; Neuburger & Egger, 2021). In tourism, the perception of risk arises in situations in which it clearly predicts the decision to avoid traveling to certain destinations; for example, due to natural disasters, terrorism, political instability or health risks (Qiu et al., 2020). Fuchs and Reichel (2006) define the perception of risk in tourism as the potential danger associated with travel. Risk perception can influence an individual's behavior. Factors that affect the perception of tourism risk include subjective and objective factors (Cui et al., 2016; Kozak et al., 2007; Reisinger & Mavondo, 2006). Subjective factors can be divided into demographic factors, such as age, gender, educational experience, academic background, social status, geography, educational level, income and social experience and individual cognitive abilities, emotional characteristics, and experiential tradition and the sociocultural paradigm (for example, the social amplification of risk, cultural theory, trust and values). The perception of risk is a central element in the decision-making process of the traveler or tourist (Kozak et al., 2007; O’Connor & Assaker, 2022) and can even alter rational decisions regarding travel or choice of destination (Sönmez & Graefe, 1998). Dryhurst et al. (2020) measure the perception of risk associated with the COVID-19 pandemic. The results confirmed the significant association between COVID-19 risks perceived by the public and their likelihood of adopting protective and preventive measures, thus affirming the role of risk perception in influencing people's behavior.

The objective of this paper is to identify the subjective and objective factors that are relevant to determine segmentation profiles of tourists, considering their future intentions regarding vacation decisions and risk perceptions and also their sociodemographic characteristics. The structure of groups detected will be useful in predicting future tourism demand and developing recovery strategies for the tourism sector.

One of the main contributions of this work with respect to previous studies (Brida et al., 2021), is that it deepens the identification of groups of different behaviors in the face of a risk situation, specifically the COVID 19 pandemic. This facilitates a more profound and nuanced design of differentiated sectoral recovery policies that allow better identification of the resilience capacities of tourist destinations, according to the profile of the tourist who visits them in terms of different types of risks.

The remainder of the document is organized as follows: section 2 presents the details of data collection, descriptive statistics and describes the methodology used for the analysis. In section 3 the empirical results are discussed. Finally, in section 4, conclusions, lessons learned, and future reflections are discussed.

Data and methodology

Questionnaire and data collection

The questionnaire applied was designed based on a previous one by the Institute of Tourism of the HES-SO Valais-Wallis, Switzerland. The questionnaire was planned using the software Sphinx iQ2 (https://en.lesphinx-developpement.fr) that allows the collection online of dynamic and interactive questionnaires that engage respondents. It was prepared in the Spanish language and distributed through emails during the period 14 Dec 2020–7 Jan 2021. The questionnaire was implemented online, and it was aimed at domestic and foreign tourists of Uruguay. The questionnaire was distributed by Uruguayan travel agencies and public and private tourist organizations to their contacts and clients, regular customers over 18 years of age, who participated voluntarily. Respondents of the study consisted of tourists who have vacationed in Uruguay or have shown interest in vacationing there. After discarding those questionnaires with incomplete or incoherent answers, three hundred and sixty-nine (369) responses were obtained. As a result, a convenience sample was obtained; the data available allowed characterizing the people who answered and analyzing the changes in their travel behavior. The sample design does not allow for expansions or extrapolations to a broader set of people than those surveyed.

The questionnaire consists of 34 questions, distributed in five sections that are detailed below:

1. Perceptions of the pandemic situation on travel behavior and risk perception. Respondents are invited to give information about their appreciation of the situation generated by COVID-19, their travel habits, and government policies. The 5-point Likert Scale was used in measuring expressions within the measurement model (ranging from 1- ‘Strongly Disagree’, 2- ‘Disagree’, 3- ‘Neither Agree nor Disagree’, to 4- ‘Agree’, 5- ‘Strongly Agree’).

2. Travel habits in a normal situation. The aim is to collect information about usual vacation spots, types of accommodation and reasons for traveling, as well as the number of trips per year in normal situations.

3. Impact of the pandemic on current travel plans or behaviors. Questions about which health safety measures in the places of destination cause the trip to be postponed or replanned due to the COVID-19 situation. This section also includes other issues about vacations such as places to visit, type of accommodation, and duration of the trip and website inquiries. In addition, the questionnaire provides a set of images based on Image wall (Boughzala & Moscarola, 2013; Boughzala et al., 2014) whose aim is to encourage open answers in order to test preferences for different types of vacations in pandemic situations. Due to space limitations, the present research does not include analysis of these data.

4. Pandemic impact on tourists’ economic situation and spending propensity. The aim is to find out whether the crisis and related restrictions have changed people's ability to save, i.e. the amount of money they have set aside, compared to a normal situation, and travel decisions.

5. Social and professional information. This part of the survey includes demographic information, such as age, gender, education level, nationality, employment status, and household income.

Descriptive analysis of respondents

Tables 1–4 summarize the main characteristics of the respondents of the questionnaire survey. The demographic characteristics of the respondents are summarized in Table 1. Out of the 369 responses received, the majority (approximately 95.39%) of the respondents were from Uruguay. Approximately 73% were female. Regarding educational level, 89.7% of those surveyed have higher education, of which 62.3% have tertiary education, 19.2% have a master's degree, and 7.9%, a doctorate. Of those who answered the survey, 44.8% were between 26 and 45 years old and 38.8% fell into the 46–65 age range. According to the Statistic National Institute (INE) and Minister of Tourism (MINTUR) survey of internal tourism in Uruguay, 30% of the resident population who are stakeholders in the domestic tourism market in Uruguay are between 25 and 44 years old, and 35% are over 45 years old. In respect to employment, 43.6% are employes, 24.4% are independent professionals and 15.4% are retired. Almost 31.8% of the respondents belong to households with a gross monthly income between USD500 and USD 1,500 and 29% of the respondents belong to households that fell into the USD 1,500 - USD 2,500 income range.

Table 1. Demographic information of the sample.

Variable	Categories	%
Age	18–25	5.42
	26–35	22.76
	36–45	21.95
	46–55	20.60
	56–65	18.16
	66–75	9.49
	76 y +	1.63
Gender	Female	72.90
	Male	26.83
	Other	0.27
Country of residence	Argentina	1.63
	Other	2.98
	Uruguay	95.39
Education level	PhD	7.86
	Primary school	0.27
	Master	19.24
	Bachelors	10.30
	University/tertiary	62.33
Employment Status	Unemployed	3.25
	Senior Executive / Officer	9.21
	Employee / Officer	43.63
	Student / Military Service	1.36
	Male/housewife (exclusively)	0.27
	Freelance / Professional	24.39
	Retired	15.45
	Skilled laborer	0.54
	Other	1.90
Monthly household income (USD)	USD 1001 to USD 1500	14.79
	USD 1501 to USD 2000	15.34
	USD 2001 to USD 2500	13.70
	USD 2501 to USD 3000	8.49
	USD 3001 to USD 3500	4.66
	USD 3501 to USD 4000	4.11
	USD 501 to USD 1000	16.99
	More than USD 4000	6.58
	Does not respond	8.22
	USD 500 or less	7.12

Source: Own elaboration

Table 2. Overall impact of COVID-19 (% of sample).

Questions	Totally agree	Agree	Neither agree nor disagree	Disagree	Strongly disagree
The current coronavirus situation worries me.	50.14	42.01	4.07	2.71	1.08
Coronavirus is just a new form of influenza.	2.98	14.09	22.22	37.40	23.31
I think that what is currently happening concerning coronavirus is very alarmist.	8.40	21.14	22.49	32.52	15.45
Coronavirus is totally out of control.	6.23	26.83	28.18	34.42	4.34
Tourism is largely responsible for the spread of coronavirus.	3.79	22.49	23.04	37.67	13.01
Tourism will be massively affected by the coronavirus shortly.	36.04	51.49	7.32	4.07	1.08

Source: Own elaboration

Table 3. Impact of control measures on planning a vacation (% of sample).

If your travel were subject to the following measures, could you give up?	I would not give up at all	Maybe not	Depends	Maybe yes	I would definitely give up
Requirement to fill out a health form	68.56	10.84	12.47	5.96	2.17
Requirement to submit a recent negative Covid-19 test	41.73	18.16	14.91	17.07	8.13
Quarantine upon arrival in destination country	18.43	9.49	18.70	21.68	31.71
Requirement to carry out a health exam in the destination country	23.85	14.36	21.41	22.22	18.16
Obligation to be vaccinated (as soon as a vaccination is available) and to show your vaccination record	29.81	11.92	20.33	16.53	21.41

Source: Own elaboration

Table 4. Perception of the pandemic situation on travel behavior and risk perception (% of sample).

Do the following changes lead you to plan a new vacation?	No way	Maybe not	Depends	Maybe yes	Totally
The end of lockdown	15.72	8.67	25.47	24.66	25.47
Reopening of all stores	40.38	17.62	19.78	13.82	8.40
Restaurant reopening	30.62	15.72	21.41	21.14	11.11
The end of social distancing measures	19.78	10.84	18.97	26.83	23.58
Decreasing restrictions on international travel	17.34	8.94	20.05	22.49	31.17
Restoration of international flights	17.34	10.57	18.43	21.41	32.25
Changes matter little because I will not plan vacations this year in any way	38.75	7.86	28.18	9.76	15.45

Source: Own elaboration

In terms of general risk perception, the vast majority of the respondents (92%) declare that they are worried about the coronavirus situation. Only 17% of respondents agree that COVID-19 is just a new form of influenza. About 26% mentioned that tourism is one of the factors responsible for the spread of COVID-19. More than 87% of those surveyed agree that tourism will be massively affected by COVID-19 in the near future.

Regarding the traumatic aftermath of the pandemic, 68.6% of respondents believe that COVID-19 is likely to have an impact on their travel habits. Currently, most of them avoid traveling by public transport within their country of residence and traveling by boat and/or plane to a different country. It should be noted that regulations on domestic public transport (public buses) in force at the time of the survey left few options for transportation.

At the time the survey was conducted the desire to travel was high. Respondents were asked to rate their desire to go on vacation on a scale ranging from 0, which means no desire, to 100, which means high desire. The mean level was 65.3 and the median 75, a high level that reflects a positive attitude to vacationing.

Respecting vacation plans, almost 33.4% of respondents were planning vacations and 3.5% were already on vacation. It is noteworthy that 57.5% of those surveyed did not answer the question about vacation planning. This could reflect a high level of uncertainty regarding upcoming holidays. Usually, the majority of Uruguayans take their vacations between the end of December and February. Only 7.6% answered they were not planning a vacation soon.

From the point of view of health and sanitary barriers, almost 53.5% of respondents would probably cancel their trip if quarantine was a measure upon arrival in the destination country. Almost 79% of the respondents would not change their planning of travel if completing a health form is a requirement, and 60% would act similarly if a negative coronavirus test at the arrival point is required.

In regard to the situations that may change the decision of not traveling in the near future, 38.75% of respondents ensured that any changes that occurred were irrelevant because they would not be taking vacations this year. On the other hand, 31% of them would change their decision of not vacationing if there was a decrease in international travel restrictions (end of border controls, quarantine, etc.) and 32% of respondents if there was restoration of international flights.

Cluster analysis

A cluster analysis was used as the methodological strategy, in order to determine homogeneous groups that may have common characteristics and, in turn, seeking to make the groups as heterogeneous as possible across them (Chatfield & Collins, 2018). The cluster (or conglomerate) analysis tool is a classification technique applicable to the field of multivariate information, which seeks to find groupings in a set of information on certain individuals.

In the present analysis, hierarchical-aggregative classification methods are used, so as the groups obtained at a certain distance level include groups obtained at a lower distance level (Chatfield & Collins, 2018). On the one hand, given the aggregative characteristic of these methods, at the beginning of the analysis, there are as many observations as there are groups (I groups) and the latter are integrated by adding observations at each stage. At each step, the closest pair of objects (those with the shortest distance) are joined to form a group. Consequently, in each step the number of groups decreases by one. In the last stage, there is a single group with I individual. On the other hand, as they are hierarchical algorithms, the groups constituted at a distance level are not separated at higher distance levels, as is done by non-hierarchical methods, which allow the reassignment of individuals in the groups in successive stages of the analysis.

A cluster analysis was carried out in order to identify homogeneous groups of individuals regarding decision-making on whether or not to take the planned vacations, given certain COVID-19 control measures or in case of changes in the health situation. For this, those variables with the greatest explanatory power of the survey were used. These variables refer to the decision of the tourist to continue or suspend their vacations under different situations of control of COVID 19 by the authorities when entering a destination, or changes in the health risk. Thus, variables representative of those control measures that may influence vacation planning by the respondents, as well as the renunciation of making a future trip were selected. Sociodemographic variables were also considered (age, sex, country of residence, educational level, profession and household income).

The Euclidean distance and the Ward's classification algorithm were used to define the clusters. The rules of detention used in the analysis were the Pseudo-F or Calinski's rule and the Pseudo-t-o Duda-Hart test. Both stopping rules suggest a four-group structure, whose main characteristics are shown in Table 5.

Table 5. Distribution of clusters, in brackets the percentage of the mode for each variable.

	Variable	Custer 1: ‘the reflexible’	Cluster 2: ‘the cautious’	Cluster 3 ‘the risky’	Cluster 4 ‘the anxious’
Socio professional information	Average age	over 45 (65.45) over 66 (12.73)	under 45 (61.84) under 25 (9.21)	Mainly middle age	36–65 (70%)
	Gender	Female (76)	Female (83)	Female (66)	Female (73)
	Country of residence	Non-residents (9.09)	Non-residents (2.63)	Non-residents (3.42)	Non-residents (5.43)
	Education	Master(30.91) High School (12.73)	University/tertiary (73.68)	University/tertiary (66) Master(19)	PhD (13) University/tertiary (59)
	Employment Status	Unemployed (5.54) Executive (20) Independent (27)	Mainly employes & independent	Mainly employes, independent & retirees	Mainly employes & independent
Assessment current situation	Worried about COVID	Totally agree (71)	Totally agree (62)	Agree (60)	Totally agree (59)
	Similar to influenza	Totally disagree (55)	Totally disagree (37)	Disagree (44)	Disagree (39)
	Very alarmist	Totally disagree (47)	Totally disagree (43)	Disagree (36)	Disagree (25)
	Out of control	Agree (29)	Agree (38)	Disagree (43)	Disagree (37)
	Tourism responsible	Disagree (29)	Disagree (32)	Disagree (47)	Disagree (33)
	Tourism affected	Disagree (51)	Disagree (54)	Agree (66)	Totally agree (64)
Planning new vacation	End lockdown	not give up at all (45)	Depends (41)	Maybe yes (38)	definitely give up (59)
	Reopening stores	not give up at all (76)	not give up at all (68)	not give up at all (31)	definitely give up (27)
	Reopening restaurants	not give up at all (49)	not give up at all (62)	Depends (31)	Maybe yes (36) definitely give up (36)
	End social distancing measures	not give up at all (44)	not give up at all (46)	Maybe yes (36)	definitely give up (51)
	Decrease travel restrictions	not give up at all (49)	not give up at all (45)	definitely give up (40)	definitely give up (59)
	Restoration flights	not give up at all (45)	not give up at all (46)	definitely give up (40)	definitely give up (59)
	Not planning vacation	not give up at all (36)	definitely give up (34)	definitely give up (43)	not give up at all (49)
Give up traveling	Health form	not give up at all (73)	not give up at all (71)	not give up at all (58)	not give up at all (82)
	Negative COVID test	not give up at all (65)	not give up at all (31)	not give up at all (29)	not give up at all (57)
	Quarantine	definitely give up (33)	definitely give up (37)	Maybe yes (32)	definitely give up (34)
	Health exam	not give up at all (40)	not give up at all (24)	Maybe yes (29)	not give up at all (30)
	Vaccines	not give up at all (29) definitely give up (29)	definitely give up (33)	not give up at all (31)	not give up at all (43)

Source: Own elaboration.

The main characteristics of the members of the clusters identified are depicted below.

Cluster 1: ‘The reflexible’. It comprises 14.91% of the sample. It is a relatively older group, 65.45% of the group is over 45 years old, and is the group with the highest percentage (12.73%) of people over 66 years of age. In this group, female (76.36%) and male (23.64%) participation remains in a similar proportion as that of the sample. This is the group with the highest proportion of non-residents (9.09%) that is relatively higher than in the sample (4.61%). This is the group with the highest percentage of people with a master's degree (30.91%) and is also the group with the highest percentage of people finishing high school at the highest educational level (12.73%). The group is characterized by having the highest percentage of executives (20%) relatively much higher than the representation of the sample (9.21%). Furthermore, the percentage of unemployed (5.54%) is higher than in the sample (3.2%). It is also the group with the highest percentage of independent workers (27%). Despite being the oldest group, it does not have a greater number of retirees.

In regard to the perception of risk related to the pandemic situation, the vast majority of this group, like the others, are worried about the coronavirus situation (90%). They consider this virus is not a form of influenza. However, it is a group with a relatively higher percentage of people who declare not to be worried about the situation (7%) compared to the sample (3.78%).

Responding to the perception of risk related to tourism in a pandemic situation, this group has a moderate behavior in respect to the precautions to take in case of traveling. 80% avoids traveling by plane or boat, 47% avoids public transport and 69% avoids traveling in their own vehicle. 72% would change their travel habits due to the pandemic.

Regarding the measures required at the destination, 72.73% would not give up traveling if they are required to complete a health form upon entering the country and 65.45% would not give up traveling if a negative coronavirus test was required to enter. This group would not give up traveling if mandatory quarantine was required (29%)-relatively more than the sample (18%). With regard to cancelation, almost 40% would not cancel traveling if a health exam was required, that is also higher than the sample values. In terms of mandatory vaccination, group members behave differently with 29% willing to cancel the trip, a percentage slightly higher than the value of the sample (21%). It should be clarified that at the time the survey was carried out, the development of vaccines was in its incipient stage and they were not yet being implemented in the country.

When asked about the situations that may change the decision not to travel in the near future, 32.73% said that any changes that occurred were irrelevant because they would not be taking vacations this year. On the other hand, if health conditions and restrictive measures changed, 38.18% declared that they could change their decisions not to travel this year. Although there could be some changes regarding the future conditions of the pandemic, which would allow controls to be relaxed, such as the reduction of social distancing or the reopening of shops and restaurants, as well as the reduction of international controls, this group would not change their decisions made in respect to vacations.

Cluster 2: ‘The cautious’. This group represents 20.60% of the sample. It is a relatively younger group in the sample with 61.84% of the group being under 45 years old while 9.21% is under 25 years old. In this group, female participation (82.89%) is higher than the proportion of the sample (72.9%). It has the highest proportion of residents (97%) and most of them have university or tertiary education (73.68%). It is a group with a labor preponderance of employes and independent workers.

When evaluating the risk of the pandemic, 92% are worried about the situation. The group also does not regard the virus as a form of influenza (56.58%) and that it is not an alarmist situation (46.05%). This group has a high percentage of people who do not define themselves with respect to these statements, choosing the option of ‘neither agree nor disagree’ with the statements about whether the coronavirus worries them, if it is a new form of influenza, or if the virus is totally out of control.

Regarding the precautionary measures to carry out with tourist activities, this is the most cautious group. The group has the highest percentages of people who agree to avoid traveling by airplanes or boats and to avoid traveling by public or their own transport. It is also the group that mostly recognizes that they are going to change their travel habits because of the coronavirus (80%).

In relation to the measures required to enter a destination, the majority of the group is willing to renounce the trip in the event that quarantine is required upon arrival (57.89%). But with regard to other health measures, such as the requirement to be vaccinated, or to present a negative certificate, or have a health analysis performed, the members express a high percentage of doubt, and declare that it would depend on the conditions.

When asked about the situations that may change the decision not to travel in the near future, almost half (44.74%) express that they will not plan vacations in this year in any way, and it does not matter if the health situation changes. Decisions not to vacation will not change even if borders, restaurants or international travel are reopened. This is the most immovable group regarding the decisions made in relation to not planning vacations in the year.

Cluster 3: ‘The risky’. This is the largest group identified, comprising 42.28% of the sample. In this group there is a more uniform age distribution, prevailing in the middle ages. Male participation (31.41%) is proportionally higher compared to the sample (26.83%). It is a group with a preponderance of employes and independent workers, and there is also a high percentage of retirees. This group has a perception of the risk of the pandemic lower than the average of the other group (85.2%), and has a higher percentage, compared to the other groups, of people who think that the attitude towards the virus is very alarmist.

In this group, there is a lower percentage of people who agree to take precautions during travel, such as not taking public transport or avoiding traveling by plane or boat. In addition, it is characterized by being the group with the greatest reluctance to accept sanitary and health measures when visiting a destination. In this sense, it shows a high percentage of refusal to travel if a health form is required when entering, a negative coronavirus test or mandatory quarantine. However, it is not the group that most rejects the requirement of a vaccination certificate.

Regarding changes in the situation which may modify their plans not to take vacations in the year, this group has a high proportion of people who are willing to reconsider the plans if the borders and restaurants are reopened, if international travel and social distancing restrictions are removed, although not at the cluster 4 level. Furthermore, 50.64% would change the decision not to travel if the health situation changes.

Cluster 4: ‘The anxious’. This group represents 24.93% of the sample. It is a middle-aged group, with 69.57% of the group between 36 and 65 years old. In this group, gender participation is similar to the proportion of the sample. Residents’ participation represents 94.7% of the group, similar to the sample. The percentage of people with doctorate degrees is the highest of all the clusters. It is a group with a labor preponderance of employes and professional/freelancers. The risk perception related to the pandemic situation is high: 94.57% agree or strongly agree that the coronavirus worries them. More than 90% consider that tourism will be massively affected by the coronavirus in the near future.

Regarding the perception of risk related to tourism in a pandemic situation, this group has the highest percentage of people who consider their travel habits will probably not change due to the coronavirus.

This group shows a greater willingness to travel, and they are the least willing to give up the trip even if certain COVID control measures are required at the destination. 85.87% would in no way give up traveling, if they were required to complete a health form upon entering the country, 68.48% would in no way give up traveling if a negative coronavirus test is required to enter and 44.57% would in no way cancel if a health examination was required. Furthermore, it is also the group that most accepts the requirement to be vaccinated, 52.17% would not cancel due to this requirement. When asked about the situations that may change the decision not to travel in the near future, 54% would change the decision not to travel if the health situation changes. Moreover, more than 80% would change their decisions not to take vacations if there were a decrease in social distance restrictions, a decrease in international travel restrictions or if international flights were restored.

Conclusions and discussion

The objective of this study was to identify groups of tourists based on their perceptions of the risk of the COVID-19 virus in undertaking tourist activities along with changes in the behavior of tourists in the face of a pandemic. The study undertook a cluster analysis, classifying tourists based on their perception of travel risk during a pandemic and its influence on changing their travel behavior. The cluster analysis found four distinctive clusters, which will provide valuable insights into consumer behavior and reactions to a pandemic situation, and to different risk control measures.

The clustering of the participants in the survey demonstrated that, despite the fact that the vast majority of the respondents had a high awareness of the risk of COVID 19, there are differentiated behaviors of the respondents regarding the decisions to travel and the health measures to be taken. We have seen that, depending on the characteristics (age, sex, educational level, etc.) people behave differently when faced with risk. Depending on these characteristics, we find groups more willing to take risks and accept to travel in times of pandemic, compared to others that are more cautious and deliberative about taking risks when traveling. Acceptance of health and sanitary control measures at the entrance of destinations also varies according to the groups. Based on the results, several implications can be provided for the tourism industry, which requires clear health and safety strategies and a reduction in the perception of risk, in order to recover its level of business activity once the threat of COVID-19 diminishes.

There are several limitations that must be considered when interpreting the results of this study. Firstly, the sample is one of convenience, which does not allow statistical inference from the results. Secondly, the findings of this study are limited by its focus on potential tourists from Uruguay. Another limitation of the study is the special period in which the survey was administered, which coincided with the recent application of measures by the Uruguayan government to cope with the pandemic and which could bias the analysis. One characteristic that must be considered is that during the implementation of the survey, there were no vaccines available in Uruguay.

Possible future lines of research may be directed to longitudinal studies or data collection at multiple points in time during a pandemic outbreak. It may also be interesting to compare the results of this study with other cultural and geographic regions. In addition, it is envisaged that future research will include regression analysis, so as to explain the factors behind the planning of vacations after the end of lockdown and/or the relaxing of measures of social distancing, etc. Additionally, it is believed that it would be of interest to compare the results of this study with those obtained with the application of the questionnaire to tourists in Switzerland, thereby adding a comparative aspect of risk analysis in the face of the pandemic. This would contribute to a better understanding of the long-term effects of infectious diseases on changing risk perception and travel behavior over time. A final thought for another possible research stream would be to investigate the effect of the pandemic on expectations and sentiments and how they affect tourism consumption decisions, especially if there are changes in long-term consumption patterns.

Regarding the design of policies, based on the results, it may be observed that the tourism industry requires health and safety policies and a reduction in the perception of risk, to recapture its activity once the threat of COVID-19 decreases. But it is also possible that not everything will return to normal once the pandemic is over. Profound and prolonged experiences, like the COVID-19 pandemic, have the potential to create new habits and produce lasting changes in behavior. Consumption habits may vary, and this may result in fundamental changes that affect companies and employes in the sector. Faced with this possibility, it is necessary to define policies that anticipate changes in tourism demand, that may seek to protect the most vulnerable companies and help them develop strategies to adapt to the new forms of tourism models and that train their workers in the new requirements of the industry.

Another line of policy must be oriented towards improvements in the health and safety of destinations. Sanitary conditions at the host country will surely be a major concern that travelers will take into account. Likewise, the sanitary control measures, which were identified as one of the barriers to the desire to travel, should be designed so that they are not seen as invasive and costly to the traveler.

Acknowledgements

A preliminary version of this paper was presented at the GAET seminar on 2 November 2021. The authors would like to acknowledge the many valuable suggestions made by the participants of this event. The authors want to thank Professor Colin Johnson (San Francisco State University: San Francisco, CA, US) for proofreading the article and Mr Thomas le Forestier, Director of the Sphinx Iberoamérica (www.lesphinx.es) for supporting this research.

Disclosure statement

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

Additional information

Funding

This work was supported by CSIC-UDELAR (project Grupo de investigacion en Dinamica Economica’; ID 881928) and Institute Tourism - School of Management- HES-SO Valais/Wallis.