Social Distance with Tourists in U.S. Counties with the Highest Historical Numbers of Reported COVID-19 Cases

Abstract

With destinations steadily ‘opening back up for business’ (while COVID-19 cases are still high in many areas), there is an increasing need to consider residents. Integrating the cognitive appraisal theory and the affect theory of exchange, this work tests a structural model examining the degree to which residents’ perceptions of COVID-19 precautionary measures explain emotions directed toward visitors, and ultimately their willingness to engage in shared behaviors with tourists. Data were collected from 530 residents in 25 U.S. counties with the highest percentages of historical COVID-19 cases per population. A total of 10 of the 12 tested hypotheses were significant, contributing to 60% and 85% of the variance explained in contending and accommodating emotions, and 53% and 50% of the variance explained in engaging in less intimate–distal and more intimate–proximal behaviors with tourists. The implications highlight the complementary use of the two frameworks in explaining residents’ preference for engagement in less intimate–distal interactions with tourists.

1. Introduction

Since the deadly impact of the SARS-CoV-2 (or COVID-19) virus was first felt worldwide, the WHO [1] has identified 28 variants of the virus (two current variants of interest; three previous variants of interest; eight previously circulating variants of interest; two variants under monitoring; and 13 formerly monitored variants). In its wake (at the time of writing, 6 October 2022), the virus has been responsible for 620,196,243 cases globally, taking the lives of 6,553,793 individuals [2]. It goes without saying that the number of cases and fatalities ebbs and flows over time, across continents, countries, regions, states/provinces, counties/cantons, cities/towns, and communities. Arguably, the riskiest places to be found are in densely populated areas where individuals have the greatest potential for contact with others [3].

Though the United States (U.S.) is relatively large in land mass compared to other countries, it is a prime example of a country with a high percentage of its residents living in proximity. According to Dobis et al. [4], 86% of the U.S. population (285 million of the 331 million overall) resides in urban counties. Many of these counties are home to some of the most visited destinations in the U.S., namely Orlando (Florida), New York City (New York), Los Angeles (California), Miami (Florida), etc. At the same time, these counties have some of the highest historical numbers of reported COVID-19 cases. At the time of writing, 96,590,395 cases have been documented in the U.S., resulting in 1,061,883 fatalities [5]. As many of these counties (and destinations within them) consider relaxing COVID-19 precautionary measures, it is crucial to gauge how comfortable residents of these populous areas feel about tourists in their communities.

Over the last two years, we have focused little attention on considering residents and their perspectives on having tourists in their communities during the pandemic. The research that has been conducted tends to focus on one individual destination [6,7,8], without necessarily considering the historical number of COVID-19 cases locally. The present work seeks to address these gaps by providing a more robust perspective on residents throughout the U.S. who live in counties with the highest percentages of reported COVID-19 cases to date. In so doing, we propose and test a conceptual model (informed by the cognitive appraisal theory and the affect theory of exchange) to determine the extent to which COVID-19 precautionary considerations inform residents’ animosity toward potential tourists in their community, ultimately to determine how such animosity translates into residents’ level of comfort in engaging with tourists (i.e., social distance). The results of this study aim to contribute to the cognitive appraisal theory and the affect theory of exchange in the context of the COVID-19 pandemic, broadening our understanding of the salient constructs contributing to residents’ social distancing from tourists. Additionally, this work seeks to inform tourism planning at the county level within the U.S. by way of the inclusion of residents’ perspectives.

2. Literature Review

2.1. COVID-19 Precautionary Considerations and Residents’ Animosity toward Tourists

Cognitive appraisal theory: Derived from the study of emotions within social psychology, the cognitive appraisal theory (CAT) was first advanced in the work of Arnold [9]. To further our knowledge of emotions, Smith and Ellsworth [10] echoed the notion that emotions are not solely explained through a categorical or dimensional approach but more accurately as a product of cognitively appraising the context in which individuals find themselves. To best understand primary emotions, Smith and Ellsworth argued that cognitions such as pleasantness, anticipated effort, certainty, attentional activity, responsibility, and control are pivotal [10]. Succinctly put, the CAT highlights that emotions are the mental states formed through the processing of personally pertinent knowledge [11]. Through this theory, we are better equipped to understand why emotional responses vary between individuals—largely based on how they appraise contextual information [12]. Appraisal processes such as assessing centrality (i.e., perceived importance), controllability, agency, fairness, and pleasantness can tell us much about how individuals arrive at their perceived emotions, specifically in a tourism context [13,14]. Now that destinations are steadily relaxing conditions for entry during the COVID-19 pandemic, it is important to ascertain residents’ appraisal of visitors’ COVID-19 precautions and risks (especially in densely populated areas with high numbers of COVID-19 cases) in efforts to explain residents’ emotions associated with tourists in their communities.

Hypotheses development. As the COVID-19 virus (and its mutations) continues to evolve, its effects on tourism stakeholders vary. Toward the beginning of the pandemic, Zenker and Kock [15] suggested that this highly contagious disease may cause residents to be wary of outsiders and less accommodating to visitors than they had been in the past. Various studies have engaged in research since that time to test the relationship between perceived risk and pro-tourism behaviors in multiple countries such as Korea [6], Spain [7], Thailand [8], and Malaysia [16]. Many of these studies support the notion that the perceived risk of COVID-19 has negatively impacted residents’ embrace of the return of visitors, despite the noted potential benefits tourism would bring to their destinations.

This perceived risk of contracting COVID-19 is likely to be an emotional response for residents who fear the unknown and the potential for long-term damages brought on by COVID-19 infections. It seems likely that they will have a lower desire to have tourists visiting their homes, due to a fear of potential harm from catching one of the COVID-19 variants. This fear can lead to negative stereotypes about tourists and where they are coming from, leading to unaccommodating behaviors and the avoidance of places and activities where tourists will be located [17]. Although destination residents may perceive the risk of catching COVID, animosity may be reduced toward tourists who are taking proper precautions while visiting. Currently, no studies exist on the level of tourist precautions; however, recent research shows that support for tourism can be based on resident–tourist proximity [7]. Additionally, residents’ own intentions to travel may show a level of trust or comfort with the proposed travel restrictions (implemented by governments) on potential visitors [18] to their community. An intention to travel could make residents more accommodating to incoming tourists.

Perceived risk initially played a significant role in shaping residents’ attitudes about tourists after the pandemic began. However, at some point, between the mass rollout of vaccines and before the rise of the first COVID-19 variant (i.e., Delta), the perceived risks associated with COVID-19 decreased dramatically in the minds of destination residents [6]. This is likely due to the changing nature of the situation surrounding COVID-19 and its ever-evolving variants. Considering that perceived risk likely varies greatly from person to person based on context, when studying resident attitudes as we move toward a post-COVID-19 society, it seems prudent to understand residents’ current level of perceived risk. Prior studies regarding residents’ attitudes toward tourism during the pandemic have used the social exchange theory (SET) to understand whether the perceived risk of catching COVID-19 would outweigh the positive financial benefits of welcoming tourists. However, given SET’s focus on the transactional nature of resident/tourist relationships, the emotional component of support for tourism has been understudied [19]. Based on this review, the following hypotheses are proposed:

H₁: 

Residents’ intent to travel in the next 12 months will significantly (negatively) predict their contending emotions (as a higher level of animosity) toward potential tourists to their community.

H₂: 

Residents’ intent to travel in the next 12 months will significantly (negatively) predict their accommodating emotions (as a lower level of animosity) toward potential tourists to their community.

H₃: 

Residents’ preferred precautions (regarding COVID-19) taken by tourists will significantly (positively) predict their contending emotions (as a higher level of animosity) toward potential tourists to their community.

H₄: 

Residents’ preferred precautions (regarding COVID-19) taken by tourists will significantly (positively) predict their accommodating emotions (as a lower level of animosity) toward potential tourists to their community.

H₅: 

Residents’ perceived risk of contracting COVID-19 will significantly (positively) predict their contending emotions (as a higher level of animosity) toward potential tourists to their community.

H₆: 

Residents’ perceived risk of contracting COVID-19 will significantly (positively) predict their accommodating emotions (as a lower level of animosity) toward potential tourists to their community.

H₇: 

Residents’ fear of COVID-19 will significantly (positively) predict their contending emotions (as a higher level of animosity) toward potential tourists to their community.

H₈: 

Residents’ fear of COVID-19 will significantly (positively) predict their accommodating emotions (as a lower level of animosity) toward potential tourists to their community.

2.2. Residents’ Animosity toward Tourists and Resulting Social Distance

Affect theory of exchange: The affect theory of exchange (ATE), also with roots in social psychology, was developed by Lawler and Thye [20] based on the seminal work by Homans [21] on the social exchange theory. According to Lawler and Thye:

“Structural interdependencies among actors produce joint activities that, in turn, generate positive or negative emotions; these emotions are attributed to social units (relationships, networks, groups) under certain conditions, thereby producing stronger or weaker individual-to-collective ties; and the strength of those group ties determines collectively oriented behavior, such as providing unilateral benefits, expanding areas of collaboration, forgiving periodic opportunism, and staying in the relationship despite alternatives” [20] (pp. 237–238).

Following this, Lawler [22] indicated that “if positive global emotions are attributed to social units, they should develop stronger affective attachments to the targeted relations or structures; if negative ones are attributed to social units, weaker affective attachments should form” (p. 328). Based on these notions, our research argues that residents’ perceived emotions based on the precautionary measures taken by tourists concerning COVID-19 would significantly impact residents’ willingness to engage in specific behaviors involving tourists in their community. Though a few studies [19,23] have recently considered the ATE (in tandem with the CAT) in the context of tourism, such work has deviated from Lawler’s intentions by not focusing on social relationships between residents and tourists, instead centering on attitudes about tourism development.

Hypotheses development: As noted throughout much of the literature on sustainable tourism, it is crucial to consider residents’ attitudes about tourism and tourists; the success of tourism, in fact, is highly contingent upon such perspectives [23]. Prior to the pandemic, scholars [24,25,26,27,28,29] emphasized the importance of social distance in explaining residents’ attitudes toward tourists and tourism development. The results revealed that place attachment [24], types of activities [25], nationality [26,27,28], interaction [25], and perceptions of tourism [29] were some of the main determinants of social distance. For example, Thyne et al. [28] found that Japanese residents had a varying degree of social distance toward tourists from different countries (i.e., Taiwan, Australia, United States, and China), revealing that as the physical/cultural similarities increased between residents and tourists, the social distance decreased. Similar to this, others [26,27] found that residents were more welcoming of tourists who were more physically/culturally similar to themselves. In addition to this, other studies [25,29] revealed that greater contact and interaction between residents and tourists reduced perceived social distance. Recently, Thyne et al. [30] found that residents who were less sympathetically understanding of tourists favored less intimate–distal interactions and had stronger negative perceptions of tourism.

Understanding how emotions are influenced by the intimacy of exchanges between residents and tourists, especially during times of uncertainty, is an area of research that will have significant implications for residents’ willingness to engage in activities with tourists. Emotions experienced during tourism exchanges are not always caused by a singular interaction but rather characterized by multiple interactions with various individuals over time [31]. Considering the changes brought about to minimize the spread of COVID-19 (e.g., social distancing, mask mandates, etc.), it stands to reason that residents would appreciate having some level of separation from visitors to avoid further spreading of the virus. In fact, recent research suggests that residents who have close physical interactions perceive a higher level of risk associated with tourists in their areas [32]. Considering the lower level of the perceived risk of contracting COVID-19 from tourists participating in more distant activities, it seems likely that residents have more animosity toward tourists the closer those tourists are to their community spaces. Given this, and based on the previous findings focused on social distance, we propose the following hypotheses:

H₉: 

Residents’ contending emotions toward potential tourists to their community will significantly (negatively) predict their level of comfort in less intimate–distal activities with tourists.

H₁₀: 

Residents’ contending emotions toward potential tourists to their community will significantly (negatively) predict their level of comfort in more intimate–proximal activities with tourists.

H₁₁: 

Residents’ accommodating emotions toward potential tourists to their community will significantly (negatively) predict their level of comfort in less intimate–distal activities with tourists.

H₁₂: 

Residents’ accommodating emotions toward potential tourists to their community will significantly (negatively) predict their level of comfort in more intimate–proximal activities with tourists.

3. Methods

3.1. Sampling and Data Collection Procedures

The target population for this study was residents of the 25 largest U.S. counties with the historically highest rates of COVID-19 cases. To begin, the researchers identified the 100 most populous U.S. counties based on U.S. Census Bureau figures [33]. Then, using the COVID-19 Dashboard [5], the number of COVID-19 cases (reported since the onset of the pandemic) was identified for each of these 100 counties. At that point, the researchers calculated the number of COVID-19 cases as a percentage of each county’s population. This resulted in the top 25 counties listed in Appendix A, ranging from 20.99% to 35.47% of the selected counties’ residents having contracted COVID-19 since the pandemic began. The rationale for selecting 25 counties was to help ensure that we secured a robust sample and to make sure we had coverage across the major U.S. regions.

To reach this population, the research team identified 380 different online community groups across the 25 counties via Facebook (i.e., Groups and Pages) and Reddit. On average, this equated to 15 groups per county. These groups were chosen based on overall group size and relevance to county residents. After obtaining permission to post the survey in each of these groups from group moderators, a link to the survey hosted on Qualtrics was posted on three separate occasions, roughly two weeks apart (between March and May of 2022). This was done to maximize coverage within the groups, but also to avoid spamming the community members. As an incentive, participants were offered the opportunity to receive one of five USD 25 gift cards once they completed the questionnaire.

After five days of receiving no responses to the third posting, the survey was closed, yielding 956 total responses. Of those responses, 213 were removed because the respondents had taken less than the mean time (322 s) to complete the survey. Additionally, 98 responses that were not 100% completed were removed. Finally, 115 responses by participants not residing in the 25 counties under consideration were removed. This resulted in a final sample size of 530 participants. Appendix B provides a breakdown of the respondents across the 25 counties. Based on a confidence level of 95%, a population size of 331 million, and a confidence interval of 5%, the ideal sample size for our study would be at least 385 individuals [34]. Our final sample size exceeded this by 145 participants.

3.2. Measures and Data Analysis

Eight constructs were measured using Likert scales. The first four pertain to COVID-19 precautionary considerations. To measure intent to travel in the next 12 months, we adapted six items from the work of Rasoolimanesh et al. [18]. Five items were adapted from Kock et al. [35] to assess residents’ preferred precautions taken by tourists. To measure the perceived risk of contracting COVID-19, we adapted five items from Joo et al. [6]. The last COVID-19 consideration involved measuring fear of the virus, for which we adopted seven items from the research by Rather [36]. Residents’ animosity was measured by asking individuals to respond to four strongly negative emotional states (i.e., contending emotions) and four less negative emotional states (i.e., accommodating emotions) when they think of tourists in their communities. These eight items were adapted from Josiassen et al. [37]. Finally, residents were asked to indicate how comfortable they felt in engaging in 14 activities with tourists in their communities. These social distance items were adapted from Aleshinloye et al. [24] and Thyne et al. [30], with seven addressing less intimate–distal activities and seven concerning more intimate–proximal activities. All construct items were presented on 5 pt Likert scales of agreement (i.e., 1 = strongly disagree; 5 = strongly agree), except for the 14 social distance items. They were presented on a 5 pt Likert scale of comfortability (i.e., 1 = very uncomfortable; 5 = very comfortable). The exact item wording is found in Appendix C.

Tests of normality were performed to consider skewness and kurtosis estimates. Data for the eight constructs were then assessed for common method bias (given the singular data collection form). Finally, a two-step analytical sequence was performed to first assess the measurement model (through confirmatory factor analysis) for psychometrics and potentially identify any problematic items, and second to examine the structural model (through structural equation modeling) by testing each of the 12 proposed model hypotheses. All analyses were undertaken using AMOS v.26.

4. Results

4.1. Sample Description

Participants were asked to respond to eight distinct socio-demographic variables which serve to describe the sample overall. The respondents’ average age was 34.4 years, with the highest percentage of individuals falling within the 31–40 age category. A slight preponderance of participants (54.5%) identified as women. The median current annual household income (before taxes) fell within the USD 50,000–74,999 response category, with the highest percentage of individuals within the same category. Most of the respondents (69.6%) identified as Caucasian; African American (8.7%), Asian or Asian American (7.9%), and American Indian/Alaska Native (7.2%) were the next largest groups of respondents. Nearly one in four individuals identified as Latinx. Finally, the respondents had been living in their county for an average of 13.7 years (with the largest percentage falling in the range between 5 and 10 years).

4.2. Data Analysis

Before examining the measurement and structural models, the full dataset was subjected to normality testing to determine if skewness or kurtosis issues were present. This was done using the AMOS v.26 software. Each item within the dataset revealed skewness and kurtosis values lower than 2.0 [38], with the exception of one item, “In the next 12 months, I intend to travel for business”. As such, this item was removed from further analysis. With the knowledge that the data were collected using one source method (i.e., online questionnaire), we examined the dataset for common method bias (CMB) potential. This was done using two methods. First, all exogenous and endogenous variables were presented as clearly separate within the questionnaire [39]. Second, we submitted all 44 remaining items (across the eight constructs within the model) to an unrotated, single exploratory factor analysis [40]. No single factor was responsible for explaining more than 31.2% of the variance. Based on these two approaches, no evidence was found indicating CMB issues.

4.3. Measurement and Structural Model Analysis

To assess the measurement model with all items across the eight constructs, confirmatory factor analysis (CFA) was performed. Considering cross-loading items, error covariances, and low AVE scores [41], we found no problematic items within the measurement model and therefore did not have to remove any items. At that point, the 44 items within the measurement model revealed standardized factor loadings in excess of 0.70, which is considered acceptable [42] (

Table 1. Measurement model results.

Factor and Corresponding Item	Mean	Stnd β (t-Value)	CR	AVE
Intent to travel in next 12 months (IT) a	3.89		0.86	0.56
IT1	3.89	0.82 (N/A c)
IT2	3.96	0.70 (16.93)
IT3	3.82	0.72 (17.63)
IT4	3.74	0.68 (16.39)
IT5	4.03	0.81 (20.36)
Preferred precautions taken by tourists (PP) a	3.86		0.89	0.61
PP1	4.08	0.79 (N/A c)
PP2	4.05	0.82 (25.13)
PP3	3.66	0.74 (16.10)
PP4	3.77	0.76 (16.05)
PP5	3.76	0.70 (14.59)
Perceived risk of contracting COVID-19 (PR) a	3.14		0.90	0.65
PR1	3.10	0.80 (N/A c)
PR2	3.67	0.63 (16.91)
PR3	2.89	0.86 (22.49)
PR4	3.03	0.83 (21.42)
PR5	3.02	0.84 (20.75)
Fear of COVID-19 (FC) a	2.70		0.92	0.62
FC1	3.11	0.77 (N/A c)
FC2	3.10	0.74 (18.49)
FC3	2.96	0.79 (20.99)
FC4	2.29	0.82 (17.91)
FC5	2.68	0.77 (17.81)
FC6	2.46	0.83 (17.29)
FC7	2.29	0.79 (18.39)
Contending emotions (CE) a	2.56		0.93	0.76
CE1	2.44	0.87 (N/A c)
CE2	2.60	0.92 (30.30)
CE3	2.69	0.85 (26.20)
CE4	2.52	0.84 (25.17)
Accommodating emotions (AE) a	2.89		0.93	0.76
AE1	3.01	0.88 (N/A c)
AE2	2.74	0.89 (28.57)
AE3	2.76	0.86 (27.18)
AE4	3.06	0.83 (30.91)
Less intimate–distal (LI) b	3.62		0.95	0.75
LI1	3.44	0.86 (N/A c)
LI2	3.67	0.87 (27.11)
LI3	3.75	0.85 (25.81)
LI4	3.65	0.86 (26.49)
LI5	3.66	0.87 (27.14)
LI6	3.50	0.85 (26.01)
LI7	3.68	0.87 (27.21)
More intimate–proximal (MI) b	3.38		0.92	0.61
MI1	3.41	0.82 (N/A c)
MI2	2.98	0.71 (18.15)
MI3	2.90	0.73 (18.85)
MI4	3.72	0.74 (19.38)
MI5	3.35	0.84 (23.12)
MI6	3.54	0.81 (21.65)
MI7	3.77	0.78 (20.21)

^a All items measured on a 5-point Likert scale (1 = strongly disagree; 5 = strongly agree); ^b All items measured on a 5-pt Likert scale (1 = very uncomfortable; 5 = very comfortable); ^c In AMOS, one loading has to be fixed to 1; hence, the t-value cannot be calculated for this item. All other t-values are significant (p < 0.001 level). Notes: n = 530, χ² = 2348.98, df = 853, χ²/df = 2.75, CFI = 0.93, TLI = 0.92, IFI = 0.93, and RMSEA = 0.06.

).

The CFA results demonstrated incremental model fit indices—comparative fit index (CFI), Tucker–Lewis Index (TLI), and incremental model fit index (IFI)—greater than 0.90. These are deemed acceptable, according to Hair et al. [42]. The root mean square error of approximation (RMSEA), as an absolute fit index, was less than 0.08. Such a result indicates a good model fit [42]. Overall, the CFA model fit was χ²(n = 530) = 2348.98, df = 853, χ²/df = 2.75, CFI = 0.93, TLI = 0.92, IFI = 0.93, and RMSEA= 0.06.

To look closer at the measurement model, we next examined the psychometric properties. Each of the eight constructs displayed a composite reliability greater than 0.80, ranging from 0.86 to 0.95, along with AVE estimates in excess of 0.50, ranging from 0.56 to 0.76. Construct validities in the form of convergent and discriminant validities were also assessed. Discriminant validity (by way of Fornell and Larcker’s (1981) calculations) was established given that the square root of each construct’s AVE was greater than the inter-construct correlation (

Table 2. Discriminant validity analysis results.

Factors	CR	AVE	LI	IT	PP	PR	FC	CE	AE	MI
LI	0.95	0.75	0.86
IT	0.86	0.56	0.53	0.75
PP	0.89	0.61	−0.28	−0.10	0.78
PR	0.90	0.65	−0.61	−0.45	0.52	0.81
FC	0.92	0.62	−0.50	−0.37	0.38	0.72	0.79
CE	0.93	0.76	−0.46	−0.35	0.24	0.68	0.61	0.87
AE	0.93	0.76	−0.63	−0.47	0.38	0.82	0.74	0.76	0.87
MI	0.92	0.61	0.82	0.53	−0.30	−0.60	−0.43	−0.46	−0.62	0.78

Note: The bold diagonal elements are the square roots of the variance shared between the factors and their measures (average variance extracted). Off-diagonal elements are the correlations between factors. For discriminant validity, the diagonal elements should be larger than any other corresponding row or column entry. CR: composite reliability; AVE: average variance extracted.

). Convergent validity was also demonstrated through (1) standardized factor loadings in excess of 0.50, (2) AVEs greater than 0.50, and (3) significant t values associated with each item standardized factor loading [42].

The structural model was then examined for fit. As shown in

Table 3. Hypothesized relationships between constructs from the structural model.

Hypothesized Relationship	B	Beta (β)	t-Statistic	Supported?
H1: IT → CE	0.03	0.02	0.55 ns	No
H2: IT → AE	−0.15	−0.12	−3.69 ***	Yes
H3: PP → CE	0.35	0.30	5.54 ***	Yes
H4: PP→ AE	0.24	0.18	4.42 ***	Yes
H5: PR → CE	0.89	0.88	10.09 ***	Yes
H6: PR → AE	0.99	0.85	11.97 ***	Yes
H7: FC → CE	0.08	0.08	1.31 ns	No
H8: FC → AE	0.17	0.14	3.02 **	Yes
H9: CE → LI	−0.17	−0.18	−3.44 ***	Yes
H10: CE → MI	−0.13	−0.14	−2.51 *	Yes
H11: AE → LI	−0.71	−0.84	−14.13 ***	Yes
H12: AE → MI	−0.67	−0.80	−12.87 ***	No

Note: The fit indices are: χ²(863) = 2719.384, RMSEA = 0.06, IFI = 0.91, TLI = 0.90, and CFI = 0.91. * p < 0.05, ** p < 0.01, and *** p < 0.001; ns > 0.05. R² _SMC: CE = 0.60, AE = 0.85, LI = 0.53, and MI = 0.50.

, the proposed structural model demonstrated good fit to the data: χ²(n = 530) = 2719.384, df = 863, and χ²/df = 3.15; CFI = 0.91; TLI= 0.90; IFI = 0.91; and RMSEA = 0.06. At this point, each of the 12 hypotheses from the proposed conceptual model (Figure 1) could be examined. Overall, 10 of the 12 hypotheses were supported (Table 3). The first two hypotheses within the model examined the effect of residents’ intent to travel in the next 12 months on their two forms of animosity—contending emotions (H₁: β = 0.02, p > 0.05) and accommodating emotions (H₂: β = −0.12, p < 0.001)—as they considered tourists in their community. The former was not significant, while the latter was. The relationships between residents’ preferred precautions taken by tourists and the two emotions were examined through Hypotheses 3 and 4. Preferred precautions were a significant predictor of both contending emotions (H₃: β = 0.30, p < 0.001) and accommodating emotions (H₄: β = 0.18, p < 0.001).

The third set of hypotheses concerning COVID-19 considerations involved examining the relationships between the perceived risk of contracting COVID-19 and the two forms of emotions. As with the previous set, perceived risk significantly explained contending emotions (H₅: β = 0.88, p < 0.001) and accommodating emotions (H₆: β = 0.85, p < 0.001). These two hypotheses demonstrated the strongest relationships within the model. Next, we examined the effect fear of COVID-19 had on contending emotions (H₇: β = 0.08, p > 0.05) and accommodating emotions (H₈: β = 0.14, p < 0.01). It was found to be not significant in the former case, and significant in the latter. Ultimately, these four COVID-19 considerations explained 60% of the variance in residents’ contending emotions and 85% of the variance in accommodating emotions associated with tourists coming into their community.

The final four hypotheses within the model looked at the influence of these forms of resident animosity on the comfort level of engaging in activities with tourists (i.e., social distance). Contending emotions significantly predicted both less intimate–distal activities (H₉: β = −0.18, p < 0.001) and more intimate–proximal activities (H₁₀: β = −0.14, p < 0.05). Finally, accommodating emotions also significantly predicted both less intimate–distal activities (H₁₁: β = −0.84, p < 0.001) and more intimate–proximal activities (H₁₂: β = −0.80, p < 0.001), reflecting the second strongest set of relationships within the model. Overall, these two forms of residents’ animosity (as measured through contending and accommodating emotions) explained 53% of the variance in the comfort level with regard to engaging in less intimate–distal activities with tourists and 50% of the variance in more intimate–proximal activities.

5. Discussion

Considering the cognitive appraisal theory and the affect theory of exchange, this paper tested a structural model examining the degree to which residents’ perceptions of COVID-19 precautionary measures explain their emotions directed toward incoming visitors, and ultimately their willingness to interact with tourists. All but two of the 12 proposed model hypotheses were supported. In contrast to previous studies [18,19,43] which found that emotions were significant predictors of intentions, the current study examined this relationship in the opposite direction (intention as a predictor of emotions). For example, Zheng et al. [19] found that negative emotions result in decreasing residents’ intentional support for tourism activities. Contrary to previous findings [19,43], the results indicated that residents’ intention to travel in the next 12 months did not consistently explain their animosity toward potential tourists visiting their own community; the intention to travel was not a significant predictor of contending emotions (H₁) though it was a predictor of accommodating emotions (H₂).

H_3–6 in the model examined the relationships between residents’ perceptions of COVID-19 precautionary measures (i.e., preferred precautions taken by tourists and perceived risk of contracting COVID-19) and residents’ negative emotions (contending emotions and accommodating emotions) towards incoming tourists. The results for the three hypotheses were in line with the findings of previous studies [6,7,8,44], though not with Agyeiwaah et al. [43], who did not find a significant relationship between the perceived risk of contracting COVID-19 and negative emotions. For example, Rey-Carmona et al. [7] revealed that there was a positive relationship between residents’ perceived risk and their attitudes toward the negative impacts of tourism, while Agyeiwaah et al. [43] indicated that the perceived risk of contracting COVID-19 significantly (positively) influenced positive emotions. Moreover, the fear of COVID-19 was not a significant predictor of contending emotions (H₇). However, on the other hand, perceived risk had a significant impact on accommodating emotions (H₈), which is almost identical to Zheng et al.’s study [45] that revealed the fear of COVID-19 as a significant predictor of travel avoidance, and Luo and Lam’s study [46] demonstrating that the fear of COVID-19 positively and significantly influenced travel anxiety. Similarly, Rather [36] found that greater fear of COVID-19 resulted in more negative attitudes.

Furthermore, residents’ contending emotions toward potential tourists significantly predicted both their level of comfort in less intimate–distal activities (H₉) and more intimate–proximal activities (H₁₀) with tourists. As proposed, each hypothesis was supported. In line with these findings, previous studies [24,25,30] demonstrated that stronger positive emotions (e.g., emotional solidarity) explained individuals’ increased willingness to engage in more intimate interactions (i.e., affinity) with others. Consistent with previous studies carried out by Qiao et al. [47] showing that residents’ negative anticipated emotions lead to a lower desire to travel, and Thyne et al. [30], reporting that residents’ negative perceptions of tourism increased their perceptions of social distance (less intimate–distal and more intimate–proximal) with tourists, the present study revealed that accommodating emotions was a negative significant predictor of both degrees of intimate activities (H₁₁ and H₁₂).

6. Conclusions

6.1. Study Implications

This is a pioneering study investigating the connections between the perceived risk of contracting COVID-19, fear of COVID-19, preferred precautions taken by tourists, intention to travel, negative emotions toward tourists (i.e., contending and accommodating), and social distance (i.e., less intimate–distal and more intimate–proximal) through the integration of two theories (i.e., the cognitive appraisal theory and the affect theory of exchange). The results demonstrated which constructs contribute to residents’ negative emotions about tourists and how these emotions influence the social distance that exists between residents and tourists. Hence, the result of the present study provides multiple theoretical and practical implications for future research surrounding residents’ perspectives about tourists as we move toward a post-COVID-19 world.

In terms of theoretical implications, the perceived risk of contracting COVID-19 was the strongest predictor of not only the accommodating (low animosity) but also contending (high animosity) emotions of the residents, which was supported by the cognitive appraisal theory. In other words, residents’ negative emotions associated with tourists in their communities (accommodating and contending) were significantly (positively) explained the most by residents’ perceived risk of contracting COVID-19 (especially in densely populated areas with high numbers of COVID-19 cases). Hence, as a practical implication, destination organizations or managers could develop initiatives to create messaging that has the potential to reduce risk perceptions and ultimately allow residents to feel greater comfort in welcoming tourists. Destination planners can also use the full power of social media to promote safety and a reliable environment to reduce these risk perceptions (e.g., how well the local medical and health systems are doing at preventing the spread of infectious diseases). Similarly, tourism authorities can better communicate their contingency plan to residents so that they regard hazards as controllable and under control.

This study also demonstrates the crucial role of residents’ appraisal of visitors’ COVID-19 precautions, which mitigates residents’ accommodating (low animosity) and contending (high animosity) emotions. To improve preferred precautions taken by tourists, government officials can either adopt strict measures (e.g., checking vaccination status and/or mandating vaccinations, controlling the schedule and number of vaccinations received, and forcing visitors to be vaccinated) or create COVID-19 measures that can be easily adopted (e.g., wearing face-mask in a crowded area or while using public transportation, adhering to marked physical spaces for queuing, washing hands frequently, and not shaking hands with residents). Though every destination is unique in its number of COVID-19 cases, political leadership, physical layout, etc., DMOs and other entities charged with managing tourism need to remain flexible and adaptive in their tourism planning and management, ready to adjust in the face of changing reactive and proactive measures enforced by government officials.

Interestingly, while the fear of COVID-19 did not significantly influence contending emotions, it was a significant predictor of accommodating emotions, which was again confirmed by the cognitive appraisal theory. Considering the emotional aspect of the cognitive appraisal theory, people may experience different emotions in reaction to the same triggering event (fear of COVID-19). To decrease this fear, tourism authorities should communicate with residents, provide accurate information about the current COVID-19 situation, and inform and include them in their tourism policy and planning geared toward preventing the spread of COVID-19.

Moreover, to reduce the fear of COVID-19 as well as the perceived risk of contracting COVID-19, and increase the preferred precautions taken by tourists, the tourism industry might also encourage businesses to engage in less face-to-face interaction (i.e., ‘touchless’ tourism). For instance, hotels can provide family-only dining areas or room service for the breakfast buffet, and the travel and tourism industries can fully utilize digital technology to cut down on pointless contact. Hence, the possibility of untouchable services and tourism should be encouraged by destination managers and policymakers as a potential option for both residents and tourists looking to lower their perceptions of risk.

The results indicated that both contending and accommodating emotions are significant predictors of both less intimate–distal and more intimate–proximal activities. These results are consistent with previous studies [30,47] and are supported by the affect theory of exchange. According to the affect theory of exchange, people (e.g., residents) feel emotionally high when interactions go well, or they can be emotionally depressed when trades go poorly. Similarly, when the level of exchange activity rises, various emotions or feelings will manifest [23], which is supported by the affect theory of exchange. In other words, perceptions can be changed depending on the exchange between residents and tourists. This study shows that the social distance between residents and tourists depends on how residents feel about the tourists. According to the present study, if residents’ negative emotions (i.e., contenting or accommodating) toward tourists grow, so will their level of social distance. In other words, residents’ negative emotions toward tourists lead to lower affinity with them or a weaker welcome extended to them.

Tourism authorities should understand that the social distance between residents and tourists is directly dependent on residents’ emotions toward tourists, and indirectly on their perceived risks and fear of COVID-19. Hence, tourism authorities should develop or adopt strategies that can decrease these fears, anxiety, and panic, to mitigate negative feelings and ultimately reduce social distance. At that point, communication with residents is key (i.e., residents’ perspectives, feelings, worries, needs, and opinions about tourism development should be considered), and locals should be treated as equal partners (i.e., by including them in the tourism planning, giving them access to decision-making, educating them, and enhancing their knowledge about COVID-19) as we move toward a post-COVID-19 society.

6.2. Limitations and Future Research Opportunities

The following limitations of this study can serve as opportunities for future research: Given the difficulty of securing a sample of residents from the 25 counties, we used an online survey instrument distributed through a convenience sampling strategy. Because of this, our findings reflect limited generalizability across the counties. Should future researchers have greater resources (i.e., time and money), probability forms of sampling (e.g., stratified random sampling, cluster random sampling, etc.) should be undertaken if similar models are to be examined. At that point, however, we encourage researchers to focus on a smaller number of counties that would allow for the testing of a comparable model while enhancing the external validity of their findings.

While collecting data via Facebook and Reddit, the research team faced several hindrances that should be shared for the cognizance of future researchers. Due to the strict privacy policy of Facebook, members of the research team often found it difficult to gain admission into groups to which they requested access. Their profiles were often flagged as spam and were temporarily deactivated by Facebook. Moreover, some of the Facebook groups and Reddit pages did not allow members of the research team to post the survey instrument on their group discussion boards. This limited the reach of the survey, and further compromised the generalizability of our findings. Using an established panel marketing company to assist in data collection may be beneficial to future researchers. However, using companies such as Qualtrics can be rather costly, and, as others have pointed out, the data are secured from professional survey takers [48].

Though survey responses were collected over the course of six weeks, the results presented in this study reflect residents’ perceptions and attitudes during a relatively short period. Considering the pace of change during the COVID-19 pandemic, the perspectives of residents toward tourists are likely to evolve according to the waxing and waning of the threat or pandemic severity levels. Future research is encouraged to utilize a more longitudinal approach in examining the perspectives of residents during local or global setbacks (especially outbreak events) in the spirit of Kamata’s [49] recent work. Doing so would allow for a better understanding of how such attitudes and social engagement levels of residents toward tourists evolve over time.

Finally, though the variance explained in residents’ animosity toward and preferred social distance from tourists was robust, our model included a limited number of constructs. This was intentional in considering both the cognitive appraisal theory and the affect theory of exchange. However, the inclusion of additional constructs would no doubt aid in increasing the variance explained in each outcome construct within our model. For instance, perceived risk [6], perceptions of positive and negative tourism impacts [7], and community attachment [50] are just three additional predictor constructs that researchers may consider. In the meantime, our findings provide some initial insight into how residents living in areas with high historical COVID-19 cases still rate precautionary measures high, which in turn contributes to their animosity toward tourists and ultimately their favoring of less intimate interactions with such visitors.