Wine Economics and Policy 11(1): 89-106, 2022 Firenze University Press www.fupress.com/wep ISSN 2212-9774 (online) | ISSN 2213-3968 (print) | DOI: 10.36253/wep-11550 Wine Economics and Policy Citation: Giulia Gastaldello, Florine Livat, Luca Rossetto (2022) Does Covid scare wine travelers? Evidence from France and Italy. Wine Economics and Policy 11(1): 89-106. doi: 10.36253/wep- 11550 Copyright: © 2022 Giulia Gastaldello, Florine Livat, Luca Rossetto. This is an open access, peer-reviewed arti- cle published by Firenze University Press (http://www.fupress.com/wep) and distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distri- bution, and reproduction in any medi- um, provided the original author and source are credited. Data Availability Statement: All rel- evant data are within the paper and its Supporting Information fi les. Competing Interests: The Author(s) declare(s) no confl ict of interest. Does Covid scare wine travelers? Evidence from France and Italy Giulia Gastaldello1,*, Florine Livat2, Luca Rossetto1 1 Department of Land, Environment, Agriculture, and Forestry (Tesaf ), University of Padova, Viale dell’Università, 35020 Legnaro PD, Italy. E-mail: giulia.gastaldello.1@phd. unipd.it; luca.rossetto@unipd.it 2 Kedge Business School, Center of Excellence Food, Wine and Hospitality, 680 Cours de la Libération, 33405 Talence, France, E-mail: fl orine.livat@kedgebs.com Abstract. Tourism is sensitive to shocks, and the Covid pandemic has profoundly changed sector dynamics. Although wine tourism is primarily a form of proxim- ity tourism, the pandemic may have aff ected wine travellers behaviour and intention to go on a wine holiday. Th is exploratory study proposes a comprehensive analysis of the impact of Covid-related fear and anxiety on wine tourism intentions aft er the fi rst lockdown while jointly considering the eff ects of solidarity, situational and per- sonal involvement with wine. An online survey was delivered to a sample of 553 wine tourists from Italy and France, two major wine tourism destinations. Results highlight changes in wine travel patterns aft er the pandemic, which boosted post-lockdown wine tourism intentions. Indeed, the latter are poorly impacted by fear of contagion while it is enhanced by dedicating time to wine in lockdown (i.e., situational involvement) and by willingness to support local wine producers. Implications for sectors stakeholders are suggested. Keywords: Covid-19, Wine tourism, travel intentions, Covid phobia, involvement with wine, structural equation modelling, solidarity. 1. INTRODUCTION As past studies highlighted, tourism is vulnerable to shocks. Natural disasters like tsunamis [1], earthquakes [2] and fl oods [3] have an inevita- ble impact on tourism fl ow. In addition, the industry is aff ected by terrorism like 9/11 in the U.S. [4], [5] or the increased frequency of terrorist attacks in France from 2010 to 2017 [6], [7] and by war [8]. A global economic crisis as the Covid-19 pandemic can also impact on tourism [9]. Th e latter has indeed highlighted the susceptibility of tourism to measures implemented to coun- teract the circulation of the virus, mainly restricted mobility and social dis- tancing [10]. Being wine tourism a tourism branch, the present article aims at off ering a fi rst comprehensive analysis how the pandemic infl uences wine tourism intentions in a post-crisis context. According to the United Nations World Tourism Organization (UNWTO), international arrivals in Europe dropped by 68% between Janu- 90 Giulia Gastaldello, Florine Livat, Luca Rossetto ary and August 2020 compared to 2019, leading to the worst negative peak since the 1950s. In the past, research has shown that international tourism has been dam- aged by other health emergencies such as the Avian flu, with more significant damage on local (i.e. Asian) tour- ism [11]. Kuo et al. [12] show that the local number of cases has affected international tourists’ arrival in SARS -affected countries but not in Avian flu-affected coun- tries. A similar result was obtained by McAleer et al. [13]. Tourism in developing economies is subject to the epidemic crisis because of induced effects due to their geographical or physical proximity to the outbreak ’s source (e.g., 14 in the case of Ebola). Nevertheless, differ- ent tourist populations can react differently to epidem- ics. For instance, pregnant women or travellers of repro- ductive age travelled significantly less to Zika-affected regions after the Zika-birth defects association became well known [15]. Lastly, eradicating infectious disease risk associated with Malaria, Dengue, Yellow Fever, and Ebola could increase international tourism demand and increase tourism expenditure [16]. Due to its strong vulnerability, the tourism indus- try has become more flexible and increasingly resilient to crises. Some shocks are transitory, even if returning to pre-disaster levels can take years. The speed of recov- ery depends on the extent of the damage caused by the disaster, on the ability of tourism stakeholders to rebuild facilities and infrastructures, and on effective com- munication stating clearly that the destination is safe [17]. This is the case of Malaysia (a developing country and second destination in Asia), subjected to the Asian financial crisis, the outbreak of Avian flu and SARS, Asian tsunami, and threat of terrorism [18]. In Taiwan, visitors’ arrivals had not fully recovered 11 months after an earthquake [19]. Cultural differences play a role in the recovery of disaster-hit destinations [20]. In the path toward recovery, the destination’s attribute can also change and attract some dark tourism [21]. Shocks can lead tourists to substitute destinations [22]. However, with the Covid-19 crisis, the tourism industry faces a pandemic, i.e., a global crisis in which substituting des- tinations is not feasible because of mobility restrictions. Lastly, tourism can respond to shocks and become an engine for economic recovery [23, 24]. In such contexts, wine tourism can be seen as local tourism substituting non-local (i.e., international) tour- ism, and it can be favoured in a context of restricted mobility and fear of contagion due to uncertainty and fear of travelling abroad. Moreover, with an economic downturn, tourists might privilege short breaks instead of more extended stays. Proximity has been identified as a critical factor for the success of wine tourism [25]. Wine tourism has also been acknowledged as a substi- tute for urban tourism, as it is perceived as safer in the case of a terrorist threat [6]. Moreover, as tourism stake- holders make a claim for more sustainable practices and for the need to question the volume growth of the inter- national tourism industry in a climate change context [10], wine tourism could be a possible answer. Follow- ing the pandemic, clusters of wineries relying mostly on foreign tourism like those identified in Conegliano Val- dobbiadene area [26] can strongly benefit of these behav- iours. In this respect, it is worth understanding post- lockdown domestic wine tourism intentions. To the best of our knowledge, though, the impact of the Covid-19 pandemic on wine tourism has not yet been analysed. Therefore, the present work aims at exploring how the Covid-19 pandemic impacted wine tourism intentions both after the lockdown (ALWTINT) and in the long-run (LRWTINT), starting from the main antecedents identified by the sector’s literature such as involvement with wine (WI) while considering new negative and positive contingency factors, such as the effect of fear and anxiety towards the virus – further referred to as Covid Phobia (CPH) –, solidarity towards national winemakers (SUPLOCW) and acquired inter- est in wine during the lockdown (AQWINT), reflecting situational involvement. Changes in wine tourism travel patterns following the pandemic are also explored. Nota- bly, we focus on two major wine tourism players, Italy and France, hosting the highest number of wine tourists in Europe (14 [27] and 10 million a year, respectively). Figure 1 shows the number of overnight stays in hotels per month in both countries, which has dramatically fallen in 2020 and 2021 compared with 2019, despite a temporary recovery during summer. Indeed, although the 2020-2021 overnight stays trend is positive (+19% and +7% in 9 months for Italy and France, respective- ly), 2021 records are still remarkably lower than in 2019 (-44% in the first 9 months of 2021 for both Italy and France). The relevance of this exploratory analysis lies in its contribution to shed light on how the Covid-shock impacted on wine tourists’ travel intentions, which is key to predicting future demand developments and drafting appropriate recovery strategies. The present study is indeed among the first to assess the impact of Covid and of the lockdown on wine tourism while mod- elling positive and negative drivers together. In light of the uncertainty around the evolution of the current pan- demic as well as of its severe consequences on tourism sector, this information is strategic to tourism operators and especially to wineries for understanding how the virus impacted wine tourists’ behaviour and effectively 91Does Covid scare wine travelers? Evidence from France and Italy plan a recovery strategy. Certainly, wine tourism is an important tool for building and strengthening brand reputation [28], boosting both awareness and demand of a product [29]. Findings also provide useful information for planning communication and marketing activities in the pandemic context. The following section (section 2) provides an over- view on the main acknowledged antecedents of wine tourism intentions, as well as on context-related factors that can impact on the latter. Section 3 describes materi- als and methods, including a description of the sample, while section 4 presents the results obtained. Finally, section 5 discusses the key findings and related implica- tions for the wine tourism sector. 2. LITERATURE REVIEW To date, an extensive literature has developed on the antecedents of wine tourism intentions [30, 31, 32]. A key element characterising wine tourism research is involvement with wine (WI), which is identified as a vital driver of the intention to partake in wine tour- ism [30, 33] affecting wine tourists experiential priori- ties [30]. The advent of an extraordinary event like the Covid-19 pandemic, though, has caused radical changes in people’s known normality on multiple levels, conse- quently impacting on their behaviour. Particularly, tour- ism has been among the hardest-hit sectors due both to the strict limitations to mobility imposed by govern- ments and to the high risk of infection connected to travelling as a social activity. In his respect, people may have developed fear and anxiety toward the virus that may negatively impact travel intentions. On the other hand, the several prolonged lockdowns imposed in most countries forced people to slow down and have poten- tially more free time to explore their interests [34]. The following sections will provide an overview of the main antecedents of wine tourists behavioural intentions identified by the sector’s literature and fear and anxiety towards the novel Coronavirus. 2.1 Fear of Covid-19 and Corona-phobia Due to its disrupting effects on worldwide economies, to its ease of transmission and the life threatening nature of the Sars-CoV-2 illness, the Covid-19 outbreak prompt- ed the diffusion of fear and anxiety in human society [35, 36, 37]. The literature defines fear as an emotion caused by danger, pain or harm [35], [38], representing the aware- ness of danger [35]. Anxiety, instead, is a psychological response to fear [39]. Differently from psychological dis- comforts deriving from other extreme events such as nat- ural disasters [40], [41], or accidents [42], those induced by human-to-human transmissible diseases like Covid-19 are extensive and long-standing [43]. Therefore, a prolonged and amplified state of fear and anxiety towards a major catastrophic situation such as the current pandemic may trigger anxiety disorders defined as phobias [44]. In this respect, Arpaci et al. [44] developed a psychometric, self-report tool – the Covid Phobia Scale (C19P-S) – to diagnose what they clas- sify as corona phobia. Particularly, high values recorded by the scale detect the presence of a state of great fear and anxiety towards the virus. The C19P-S is originally composed of 4 dimensions – economic, psychological, psychosomatic and social – representing the four main domains affected by the pandemic. The social dimen- sion is particularly relevant when dealing with (wine) tourism activities since Covid-19 is an airborne disease, spread through small liquid particles, called droplets, emitted when talking, coughing or sneezing [45]. In this regard, travelling is potentially connected with a great risk of infection implying uncontrolled contact with thousands of individuals. Although the global scale of this health crisis may have levelled out the perceived risk of infection when traveling to other destinations [35], fear and anxiety towards the virus can lead to identi- fying travelling as a dangerous activity and to avoid it. Consequently, subjects manifesting greater levels of Cov- id phobia may show weaker post-lockdown wine tourism intentions (ALWTINT). 0,00 5.000.000,00 10.000.000,00 15.000.000,00 20.000.000,00 25.000.000,00 30.000.000,00 M on th Jan ua ry Fe bru ary M arc h Ap ril M ay Ju ne Ju ly Au gu st Se pte mb er Oc tob er No ve mb er France (source: INSEE) 2019 2020 2021 0,00 5.000.000,00 10.000.000,00 15.000.000,00 20.000.000,00 25.000.000,00 30.000.000,00 35.000.000,00 40.000.000,00 45.000.000,00 50.000.000,00 Jan ua ry Fe bru ary M arc h Ap ril M ay Ju ne Ju ly Au gu st Se pte mb er Oc tob er No ve mb er De ce mb er Italy (Source: Istat) 2019 2020 2021 Figure 1. Monthly overnight stays in hotels. 92 Giulia Gastaldello, Florine Livat, Luca Rossetto Hence, we postulate the following hypotheses: H1. Covid phobia (CPH) impacts negatively on post- lockdown wine tourism intentions (ALWTINT). H2. Covid phobia (CPH) mediates the effect of long- run wine tourism intentions (LRWTINT) on post-lock- down wine tourism intentions (ALWTINT). 2.2 Involvement with wine The key role of involvement in marketing is widely recognized among scholars [46] as it is acknowledged to affect consumer decision-making processes and behav- iour [47, 48]. The literature distinguishes among three types of involvement: enduring or personal, connected to the presence of a long-term personal relevance [50], [51], physical, arising from specific product character- istics, and situational, which is short-term and results from temporary changes in a consumer’s environment [49]. Personal product involvement is the most com- monly adopted and it is defined as a subject’s perceived relevance of an object based on his/her inherent needs, values, and interests [49, p.342] Considering the hedonic nature of wine and wine tourism consumption, it is not surprising to find extensive sector research embodying the concept of involvement [52, 53, 54]. Hedonic prod- ucts, indeed, tend to create higher involvement [55]. Particularly, findings reveal that product involvement can significantly affect wine consumers when choos- ing which wine to purchase [53], impacts on wine tour- ists’ behavioural intentions [30], motivations [32] and travel patterns [33]. However, the extent of its effect may change based on socio-demographics such as age [54], [56]. Since wine tourism falls into the category of leisure travel activities, the most appropriate type of involvement to be considered according to the literature is personal involvement, also referred to as ego-involve- ment. Recently, Bruwer and Huang [56, p.463] defined the concept of personal involvement in the field of wine research as “a motivational state of mind of a person with wine or wine-related activity…which reflects the extent of personal relevance of the wine-related decision to the individual in terms of one’s basic values, goals, and self-concept.” In this respect, Brown, Havitz & Getz [33] concep- tualized a tool to capture ego-involvement with wine in the wine tourism context – the Wine Involvement Scale (WIS) – by extending Laurent & Kapferer’s [57] wide- ly applied Consumer Involvement Profile (CIP) scale. Indeed, the CIP scale has been adopted by several tour- ism studies in different cultural contexts which contrib- uted to proving its consistency [58]. Notably, the WIS developed by the authors includes three dimensions: expertise, enjoyment, and symbolic centrality. By seg- menting a sample of fine wine consumers based on the wine involvement construct, the authors found that dif- ferent involvement segments show significantly different intention to visit a wine region in the near future, high- lighting the central role of involvement in predicting wine tourism. Sparks [30] further underlined the criti- cal role that ego-involvement (i.e., personal involvement) can play as a motivator in wine tourism. The following hypotheses are accordingly proposed: H3. (Personal) involvement with wine (WI) posi- tively affects post-lockdown wine tourism intentions (ALWTINT) H4. (Personal) involvement with wine (WI) positive- ly affects long-run wine tourism intentions (LRWTINT) 2.3 Acquired interest in wine and solidarity during the first lockdown As mentioned above, the high infection rate of Cov- id-19 [35] forced entire countries into lockdowns dur- ing which only first necessity industries (e.g., food and pharmaceutical industries) were allowed to operate. Obliged to slow down, people found more free time on their hands which could be dedicated to exploring their interests and to leisure activities [34]. Interest is defined as the degree of enjoyment a subject gets from engaging in specific activities [59]. Based on the literature, it can be affirmed that wine tourism is driven by an underlying interest, at various levels, in wine [33], [60]. Therefore, wine tourists have plausibly employed part of their free time engaging in wine-related activities, as some people did with cooking [60], thus reinforcing their interest in wine. Interest in wine, in its turn, is connected to the degree of involvement with the topic – i.e., to its subjec- tive relevance for the individual – [49]. Consequently, the new normality of the lockdown may have fostered a situational involvement with wine, boosting the effect of enduring involvement with the product as an anteced- ent of leisure tourism intentions [62]. As involvement is an antecedent of the decision to partake in wine tour- ism, it is reasonable to hypothesize that also situational involvement (i.e., an increased interest in wine follow- ing the lockdown) reinforces both long-term and short- term wine tourism intentions. Indeed, interests can drive intentions [59]. Moreover, it can amplify the predictive power of personal involvement with wine on the inten- tion to visit a wine region. H5. Acquired interest in wine (AQWINT) mediates the effect of involvement with wine (WI) on post-lock- down wine tourism intentions (ALWTINT). 93Does Covid scare wine travelers? Evidence from France and Italy H6. Acquired interest in wine (AQWINT) mediates the effect of involvement with wine (WI) on future wine tourism intentions (LRWTINT). H7. Acquired interest in wine (AQWINT) posi- tively affects post-lockdown wine tourism intentions (ALWTINT). H8. Acquired interest in wine (AQWINT) positively affects long-run wine tourism intentions (ALWTINT). As pointed out by other academics [63], a crisis of the proportions of Covid-19 encouraged the popu- lation to prioritize society’s problems over personal needs, pushing them to support national winemakers in their struggle to survive by purchasing their products. This sentiment is even more plausible considering that, already before the Covid-19 outbreak, the literature was stressing the relevance of wine tourism as a tool for sus- tainable rural development [64, 65], and the strong asso- ciation between direct sales of local producers and the desire to support to local communities [66]. Accordingly, direct sales are one of the pillars around which the wine tourism industry is built [67, 25, 28]. As a result, solidar- ity with national wineries is expected to be a positive antecedent of wine tourism intentions and to increase the willingness to go on a wine holiday after the lockdown. H9. Wi l l i ng ness to suppor t loc a l w i ner ies (SUPLOCW) positively affects post-lockdown wine tour- ism intentions (ALWTINT). H10. Wi l lingness to suppor t loca l w ineries (SUPLOCW) positively affects long-run wine tourism intentions (LRWTINT). 3. MATERIALS AND METHODS 3.1 Data collection and survey The population of interest for the study is Italian and French wine consumers having an interest in wine and wine tourism. Given the pandemic circumstances, an online survey was launched and diffused via e-mail and Facebook groups dealing with travel and oeno- gastronomy. Specifically, over 40 Facebook groups and wine stakeholders were involved, and shared the survey with their online communities. Data collection lasted two months, June and July 2020. Alike Villacé-Molin- ero, Fernández-Muñoz et al. [68], snowball sampling is deemed an appropriate sampling technique to explore travel intentions considering the urge to collect data on a rapidly evolving phenomenon under unprecedent- ed circumstances. This technique has been previously adopted in tourism and social science research [e.g., 69, 70], allowing to shrink time and monetary costs of data collection and to recruit hard to reach communities [71] while accounting for multiple eligibility requirements [72]. The main disadvantages of snowball sampling are self-selection bias and over-representation of subgroups having similar characteristics [72]. These limitations were addressed by collecting a large sample and by try- ing to diversify it socio-demographically. The questionnaire included 7 main sections inves- tigating the following dimensions: socio-demograph- ics, ego-involvement with wine (WI), Covid phobia (CPH), acquired interest in wine during the pandemic (AQWINT), previous wine tourism experience, wine tourism intentions (LRWTINT, ALWTINT), and finan- cial difficulties caused by the pandemic. Specifically, the socio-demographic section captured age, gender, education, country of residence, household composition, marital status, household income before the pandemic. Household income is captured through 4 descriptive statements adapted from Istat annual survey on life con- ditions. For example, A sufficient economic situation is described as follows: “My monthly household income was usually just sufficient to cover expenses and I/we could hardly save part of it.” Potential economic constraints to travel are captured through one statement measuring family income vari- ations following the pandemic, ranging from 1=much worse, to 5=much improved (Table 1). Wine tourists are identified through one statement assessing if the respondent visited a wine producing region and/or participated in a wine festival in the last 3 years [33]. Involvement with wine (WI) is captured through Brown et al.’s [33] wine involvement scale (WIS), which is deemed the most appropriate for the present study due to its solid theoretical foundation and its specific application to wine tourism studies. The original WIS includes 15 items measured on a 7-point Likert scale, where 1 = totally disagree and 7= totally agree. Fear and anxiety towards Covid (CPH) are captured by adapting Arpaci et al. [44] Covid-19 phobia scale (C19P-S). In the present study, the C19P-S is preferred to similar scales [e.g., 37] due to its capability to embody the effects of both Covid-related fear and anxiety. Con- sidering the aim of the study, which is not diagnostic but rather to highlight potential negative effects of Covid-19 on wine tourism intentions, the adapted C19P-S scale (further referred to as CPH scale in the text) includes the psychological and social dimension measured through 7 items selected based on loading scores. Like the wine involvement construct (WI), items are measured on a 7-point Likert scale, where 1 = totally 94 Giulia Gastaldello, Florine Livat, Luca Rossetto disagree and 7= totally agree. Five items measured on a 7-point Likert scale (1 = totally disagree to 7= totally agree) are introduced spe- cifically for the present study to capture the effect of the lockdown, and particularly of having more free time because of it, on interest in wine (AQWINT). Long-run wine tourism intentions (LRWTINT) are captured through a single item adapted from Sparks [30] and measuring the willingness to take a wine trip in a future holiday on a 7-point Likert scale (1 = totally disa- gree and 7= totally agree). An additional item captures the short-term intention to go on a wine trip after lifting Covid-related mobil- ity restrictions (ALWTINT) – i.e., at the end of the first lockdown – measured on a 7-point Likert scale. Finally, one item captures willingness to support local wineries by purchasing locally produced wines (SUPLOCW) on a 7-point Likert scale (1 = totally disa- gree, 7= totally agree). The item is formulated as follows: “After the COVID-19 pandemic, I think it is important to support Italian winemakers by purchasing wines pro- duced locally”. A detailed description of the items adopted for each scale and construct is provided in Table 2. 3.2 Sample description A total of 751 questionnaires was collected. Incom- plete surveys were excluded, and the final sample was reduced to 713 valid questionnaires. For the sake of the analysis, only people having previous wine tour- ism experience were considered (n=553), 412 of whom from Italy and 141 from France. Table 1 summarizes the socio-demographic profile of the sample by Country of residence of the respondents. The socio-demographic characteristics of the sample are in line with the profile of wine tourists reported by the literature, which identi- fies them as highly educated tourists aged from 30 to 50, with typically woman travelling with their partner, with a high income [61, 73, 74, 75]. Notably, both samples present similar shares of males and females while highlighting a slight prevalence of females (53.2% in Italy; 53.9% in France). Compared to France, Italy records a higher share of singles (50.5%) and a lower average education level (17.5% of post-grad- uates against the 56.0% observed for France). In both samples, most respondents enjoy either a sufficient or good economic situation before Covid-19 that did not change following the pandemic (65.0% in Italy, 66.7% in France). Nevertheless, a remarkable share of interviewees from both countries declares that his/her family income has worsened after Covid-19 (31.8% Italy; 27.0% France). 3.3. Data Analysis A preliminary descriptive analysis is conducted through SPSS software to explore wine tourism travel patterns before the pandemic, as well as wine tourism intentions after mobility bans are lifted (ALWTINT), and long-term tourism intentions (LRWTINT), among Italian and French wine tourists. AMOS software is used to further perform Structural Equation Modelling (SEM). SEM is widely applied in many fields of study dealing with human-based data, particularly in con- sumer behaviour studies, tourism included [32, 76, 77). Indeed, this methodology allows path modelling and the simultaneous estimation of measurements through mul- tiple equations. Differently from similar techniques such as Partial Least Square (PLS), SEM estimation accounts for error variance. This represents a considerable advan- tage for behavioural studies, where complex theoreti- cal concepts (such as the fear of the novel Coronavirus) cannot be measured directly through a single item. Still, instead, they are captured by multi-item latent con- structs [78]. By accounting for the measurement error associated with the use of latent constructs and correct- ing for it, SEM can provide higher robustness for elabo- rations made on data collected from human individu- als, which are often not normally distributed [78]. SEM consists of two main steps: Step 1 is the evaluation of the measurement model (MM), and step 2 is the analysis of the causal relationships between constructs, i.e., the structural model (SM) analysis. To proceed with step 1, exploratory factor analysis (EFA) and confirmatory fac- tor analysis (CFA) are run on the 3 constructs included in the MM – i.e., Covid phobia (CPH), involvement with wine (WI) and acquired interest in wine during the lockdown (AQWINT). First, the factor analysis (EFA) with principal com- ponent as the extraction method and oblique rotation is run. Like in other studies [30], oblique rotation is chosen as a correlation among the items expected. The EFA con- firms the 3 constructs load on different factors, 4 of the 6 items referring to symbolic centrality of WI scale load on a different factor showed no consistency with the rest of the scale. This is in line with past research highlight- ing potential inconsistencies of the symbolic centrality dimension of involvement as the context changes [58]. Therefore, the symbolic centrality dimension is dropped, contributing to maintain an adequate sample-size/ parameters ratio for SEM analysis [78]. Based on Cron- bach’s alpha, other items are trimmed from both CPH and WI scales. The final WI scale includes 7 items, while CPH comprise 5 items. No items are removed from AQWINT scale (5 items). 95Does Covid scare wine travelers? Evidence from France and Italy Secondly, we proceed with the confirmatory fac- tor analysis (CFA) of the measurement model (MM), the results of which are presented in Table 2. To evalu- ate MM’s Goodness-of-fit (GOF), Root Mean Square Error of Approximation (RMSEA) and Standardized Root Mean Residual (SRMR) are considered as indices of absolute fit. At the same time, Tucker Lewis Index (TLI) and Comparative Fit Index (CFI) are reported for incremental fit. Thresholds for the GOF indices are con- sidered based on sample size (n) and on the number of observed variables in the model (m) according to Hair et al.’s guidelines [78]. Overall GOF of the measurement model (MM) on the whole sample is satisfactory (χ2 (553) = 441.13; df = 112; p < .001; χ2/df = 3.94; RMSEA = .07; CFI = .96; TLI = .95; SRMR = .04). Although some researchers argue that χ2 should not be significant [e.g.., 30], this statistic tends to penalize larger samples and models with a higher number of observed vari- ables [78]. According to sample size (n = 553) and num- ber of observed variables (m = 17) of the MM applied, significant p-values for χ2 are expected [78]. Construct Reliability (CR) and Average Variance Extracted (AVE) are above the recommended thresholds for all latent constructs (CR > .7; AVE > .5) [78], and all standard- ized factor loadings are significant and above the ideal threshold (.7) providing evidence of convergent validity for all scales [78]. Discriminant validity is also support- ed by AVE exceeding inter-construct correlations [78]. For step 2, i.e., the analysis of the causal relation- ships between constructs, the same GOF indices used for the MM are considered. Mediation effects (H2; H5; H6) are explored in addition to direct effects and are estimated through bootstrapping (500 bootstrapping intervals) with bias-corrected confidence intervals (C.I. = 95%). This technique is reported to be a reliable tool to test for indirect effects, providing intervals for estimates without relying on distribution [79]. Lastly, cross-cultural differences between France and Italy are further explored through a multigroup analysis (MGA). Before path differences between the two Table 1. Socio-demographic profile of respondents by country. Italy (n=412) France (n=141) Frequency % Frequency % Gender Male 193 46.8 65 46.1 Females 219 53.2 76 53.9 Age 18-29 76 18.4 24 17.0 30-40 121 29.4 36 25.5 41-50 103 25.0 38 27.0 51-60 82 19.9 26 18.4 60+ 30 7.3 17 12.1 Education High school or lower 13 3.1 0 0.0 College 129 31.3 13 9.2 University 198 48.1 49 34.8 Post-Graduate 72 17.5 79 56.0 Marital Status Couple 204 49.5 106 75.2 Single 208 50.5 35 24.8 Has children No 329 79.9 99 70.2 Yes 83 20.1 42 29.8 Income Before Covid Insufficient 3 0.7 4 2.8 Just sufficient 35 8.5 17 12.1 Sufficient 194 47.1 71 50.3 Good 180 43.7 49 34.8 Income Variation Much worse 12 2.9 6 4.2 After Covid Worse 119 28.9 32 22.7 Unchanged 268 65.0 94 66.7 Improved 12 2.9 9 6.4 Much Improved 1 0.2 0 0.0 N=355 Source: own elaboration. 96 Giulia Gastaldello, Florine Livat, Luca Rossetto countries are tested, a preliminary multigroup confirma- tory factor analysis (MCFA) is required to test for the measurement model to be consistent between the two groups. To do so, the fitting of the MM is first tested on the French and Italian samples separately to assess con- figural invariance. The latter condition is confirmed by the MM showing acceptable fitting for both groups (Italy χ2 (412) = 361.77; df=112; p < .001; χ2/df = 3.23; RMSEA = .07; CFI = .96; TLI = .95; SRMR = .04; France χ2 (141) = 242.99; df=112; p < .001; χ2/df = 2.17; RMSEA = .09; CFI = .94; TLI = .93; SRMR = .05). Moreover, the totally free multiple group model (TF) reveals acceptable fit (χ2 (553) = 605.10; df=224; χ2/df = 2.70; p < .001; RMSEA = .05; CFI = .96; TLI = .95; SRMR = .04). All standardized factor loadings are significant at p < .001 with values of .7 or above in both groups [78], supporting configu- ral invariance. Subsequently, we test the model for met- ric invariance by comparing the fit of the constrained model (M1), where all factor loadings are imposed to be equal between the groups, and of the unconstrained model (M0), through a likelihood ratio test (LR). LR test compares the model with and without constraints by estimating them as nested models. The output produces a chi-square χ2 statistic estimated according to equation 1 [79]: 𝜒𝜒! = −2𝑙𝑙𝑙𝑙𝑙𝑙) 𝐿𝐿 (𝑀𝑀#) 𝐿𝐿 (𝑀𝑀$) . = {−2𝑙𝑙𝑙𝑙𝑙𝑙[𝐿𝐿 (𝑀𝑀#)]} − {−2log[𝐿𝐿 (𝑀𝑀$)]} (1) 𝜒𝜒! = −2𝑙𝑙𝑙𝑙𝑙𝑙) 𝐿𝐿 (𝑀𝑀#) 𝐿𝐿 (𝑀𝑀$) . = {−2𝑙𝑙𝑙𝑙𝑙𝑙[𝐿𝐿 (𝑀𝑀#)]} − {−2log[𝐿𝐿 (𝑀𝑀$)]} Table 2. Factor loadings and reliability of the measurement model. Item description Factor loadinga Average Variance extracted (%)b Construct Reliabilityc AVE CR Covid Phobia (CPH) CPH1 The fear of coming down with coronavirus makes me very anxious. 0.91 67.9 .91 CPH2 I am extremely afraid that by traveling me/ my family might become infected by the coronavirus. 0.81 CPH3 News about coronavirus-related deaths causes me great anxiety. 0.88 CPH4 After the coronavirus pandemic, I feel extremely anxious when I see people coughing. 0.76 CPH5 The idea of traveling with big groups of peolpe (e.g., by train or plane) makes me anxious 0.78 Involvement with wine (WI) WI1 My interest in wine makes me want to visit wine regions 0.80 73.9 .95 WI2 My interest in wine has been very rewarding 0.86 WI3 Wine represents a central life interest for me 0.84 WI4 Wine represents a central life interest for me 0.92 WI5 I have invested a great deal in my interest in wine 0.92 WI6 Much of my leisure time is devoted to wine-related activities 0.90 WI7 People come to me for advice about wine 0.78 Acquired Wine Interest in lockdown (AQWINT) AQWI1 During the lockdown, I learnt more about wine and winemaking 0.82 69.6 .92 AQWI2 During the lockdown, I became more passionate about wine 0.81 AQWI3 During the lockdown, I watched and/or read on-line content (e.g., youtube videos, blogs) and/ or documentaries about wine 0.87 AQWI4 Since the beginning of the lockdown, I started following profiles of wineries/wine experts on social media 0.87 AQWI5 Since the beginning of the lockdown, I started looking for more information about the wines I want to purchase 0.80 n=553. a. Based on standardized regression weights from AMOS. b. AVE was computed based on the formula from Hair et al. [78] as an indicator of convergent validity. c. CR was computed based on Hair et al. [78]. 97Does Covid scare wine travelers? Evidence from France and Italy This step brings statistical evidence that the measure- ment model (MM) measures the same constructs in both the groups considered: if the χ2 statistic between the two models is significant, it means model estimates differ between the groups. I our study, model’s metric invariance is supported (χ2 test p= .625), confirming the equivalence of psychometric properties of the MM across groups [78]. Therefore, it is appropriate to proceed with multi-group comparisons. Single paths are further tested to identify which effects are significantly different between groups. In light of the size difference between the two groups, estima- tions have been weighted over groups numerosity. 4. RESULTS 4.1 Wine tourism travel paths before Covid and post-lock- down travel intentions. Descriptive statistics of the samples are presented in Table 3. Before the pandemic, most Italian and French wine tourists travelled to wine regions close to their area of residence and/or located in different regions, and a remarkable share visited wine regions in other EU countries (34.2% in Italy; 34.8% in France). The average length of stay is slightly higher for French wine tour- ists, who tend to travel with their partner (59.6%), with friends (41.1%) or their family (29.8%), prefer private lodgings (41.1%) or hotels (34.4%) as accommodation, and declare a higher average budget compared to Ital- ian tourists. However, this budget difference is not sig- nificant (F (1, 508) = 2.26, p = .13). Instead, Italian wine tourists tend to prefer shorter trips (the 43.4 visits to a wine region no longer than 1 day), and usually stay at bed & breakfasts (38.4%) or hotels (29.3%). Similarly to French wine tourists, most Italians usually travel with their partner (55.8%) or friends (54.4%), but a consider- ably higher share travels with other wine lovers (28.9% in Italy; 17.0% in France). With respect to wine holidays after mobility restric- tions, the great majority of both French and Italian wine tourists plans wine travel in a different region and to stay Table 3. Wine tourism travel patterns before and after Covid-19. Before Covid After Covid* Italy France Italy France freq. % freq. % freq. % freq. % Visited wine regions in: The same region where I live Yes 306 74.3 88 62.4 133 41.0 29 33.3 A different region in my country Yes 292 70.9 106 75.2 241 74.4 54 62.1 Another E.U. country Yes 141 34.2 49 34.8 95 29.3 32 36.8 An Extra E.U. country Yes 34 8.3 24 17.0 20 6.2 6 6.9 Length of stay 1 day or less 178 43.4 43 30.9 75 23.1 16 18.4 2-3 days 156 38.0 57 41.0 145 44.8 29 33.3 4-7 days 65 15.9 24 17.3 62 19.1 28 32.2 ≥ 7 days 11 2.7 15 10.8 25 7.7 14 16.1 Preferred accommodation Hotel 68 29.3 33 34.7 43 18.5 22 31.0 Bed & Breakfast 89 38.4 13 13.7 89 38.4 6 8.5 Private lodging 39 16.8 39 41.1 39 16.8 37 52.1 Camping/village 9 3.9 5 5.3 8 3.4 3 4.2 Agritourism 27 11.6 5 5.3 53 22.8 3 4.2 Traveling with partner Yes 230 55.8 84 59.6 193 59.6 50 57.5 Traveling with friends Yes 224 54.4 58 41.1 157 48.5 30 34.5 Traveling with family Yes 75 18.2 42 29.8 51 15.7 24 27.6 Traveling with wine lovers Yes 118 28.6 24 17.0 57 17.6 10 11.5 Traveling alone Yes 33 8.0 13 9.2 24 7.4 8 9.2 Budget (€) 431.0 513.0 539.9 622.3 N=553: Italy n=412; France n=141. *After Covid wine travel statistics refer solely to wine tourists who are most likely to have a wine holiday after the end of mobility restric- tions (ALWTINT ≥ 4; France n = 87; Italy n = 324). 98 Giulia Gastaldello, Florine Livat, Luca Rossetto longer than one day (44.8% 2-3 days in Italy; 65.5% % 2-7 days in France). Among Italian respondents, the inter- est in hotels dropped by 58% in favour of an agriturismo (+97 %; Table 3), which are typically family run farms with a limited number of rooms. Th is variation does not seem to be related to fear and anxiety towards Covid as no signifi cant diff erence in CPH emerged for wine tour- ists preferring an agriturismo (F (1, 322) = 1.5, p = .22) or a hotel (F (1, 322) = 1.7, p = .20) for a post-lockdown wine holiday. Most French tourists still prefer private lodgings (+27%) and are interested in hotels (31.0%). Generally, the Italian sample shows a signifi cantly higher intention to go on a wine holiday both on the long-term and aft er the lift ing of mobility bans (Table 4). 4.2 Structural model results Th e structural model (SM) is fi rst tested on the whole sample (Figure 2). Goodness-of-f it statistics reveal a satisfactory fi t to the data (χ2 (553) = 605.81; df = 175; p < .001; χ2/df = 3.46; RMSEA = .07; CFI = .95; TLI = .95; SRMR = .04). Th e model shows a remarkable predictive power, explaining 41% and 42% of LRWTINT and ALWTINT variance respectively. Involvement with wine is a signifi cant antecedent of long-term wine tourism intentions (WI -> LRWTINT; β = .57; p < .001), which is the main predictor, followed by willingness to support national wineries (SUPLOCW -> LRWTINT; β = .15; p < .001). As regards the willing- Table 4. Long-term and short-term wine tourism intentions. 1 2 3 4 5 6 7 Mean St.Dev. Anova F p Would like to visit a wine region in a future holiday (LRWTINT) Italy 0.7 1.7 1.9 6.8 9.0 16.0 63.8 6.3 1.25 85.98A 0.00 France 7.1 7.8 11.3 14.9 23.4 12.1 23.4 4.7 1.85 Plans to visit a wine region aft er mobility bans are lift ed (ALWTINT) Italy 5.8 7.3 8.3 6.6 14.1 15.8 42.2 5.3 1.93 29.23 0.00 France 12.8 11.3 14.2 12.1 17.0 11.3 21,3 4.3 2.02 n=553. 1=strongly disagree; 7=strongly agree. A Th e assumption of Homogeneity of Variance is violated, Welch Anova is used. Figure 2. Path diagram with standardized regression coeffi cients: SEM results on the whole sample. Note: n = 553; ***p < .01; **p < .05; *. Signifi cant paths are represented with a continuous line and the related structural weights are reported in bold. 99Does Covid scare wine travelers? Evidence from France and Italy ness to go on a wine holiday after the lifting of mobility restrictions (ALWTINT), it is significantly predicted by both LRWTINT (β = .52; p < .001), and by AQWINT (β = .11; p = .04). A worse family income following the pandemic (WORSEINC) positively affects ALWTINT as well, although to a lesser extent (β = .09; p = .01). Interestingly, neither WI nor SUPLOCW are predictors of ALWTINT. Covid-related fear and anxiety (CPH) have a limited negative impact on post-lockdown wine tourism intentions (CPH -> ALWTINT β = - .07; p = .05) but no significant effect on LRWTINT. Finally, as expected, WI is a significant antecedent of AQWINT in lockdown (β = .75; p < .001). While the relationship between WI and LRWTINT is not significantly mediated by AQWINT, the effect of WI on ALWTINT is fully mediated by the construct (direct effect β = .07; p = .28; indirect effect β = .09; p = .04). Regarding mediation of CPH among LRWTINT and ALWTINT, a significant indirect effect was found (β = - .01; p = .04), although having a limited size. Table 5 summarizes the results obtained from the SEM analy- sis for all the hypotheses postulated while correlations, mean, and standard deviation of the variables included in the path diagram are proposed in Table 6. Multigroup comparisons between French and Italian wine tourists are conducted to check for cross-cultural differences in single paths of the model. Table 7 summa- rizes the key descriptive statistics of the two sub-samples compared through the multigroup analysis (i.e., France and Italy). The effect of AQWINT on ALWTINT is found to differ significantly between France and Italy (χ2 (351, 553) = 8.01, p < .001). In particular, the effect for Italian respondents is positive and significant (β = .20; p < .001), while it is negative and non-significant for the French sub-sample (β = - .18; p = .13). Slightly significant differ- ences are found also for the effect of CPH and of WOR- SEINC on ALWTINT (χ2 CPH (351, 553) = -.22, p = .07; χ2 WORSEINC (351, 553) = 2.65, p = .09). Similarly to the former effect, the two paths are not significant in the Table 5. Summary of hypotheses tested and related outcomes. Hypothesis Outcome H1. Covid phobia impacts negatively on post-lockdown wine tourism intentions. Partially supported H2. Covid phobia mediates the effect of future wine tourism intentions on post-lockdown wine tourism intentions. Not supported H3. Involvement with wine positively affects post-lockdown wine tourism intentions. Not supported H4. Involvement with wine positively affects future wine tourism intentions. Supported H5. Acquired interest in wine mediates the effect of involvement with wine on post-lockdown wine tourism intentions. Supported H6. Acquired interest in wine mediates the effect of involvement with wine on future wine tourism intentions. Not supported H7. Acquired interest in wine positively affects post-lockdown wine tourism intentions. Supported H8. Acquired interest in wine positively affects long-run wine tourism intentions. Not supported H9. Willingness to support local wineries positively affects post-lockdown wine tourism intentions. Not supported H10. Willingness to support local wineries positively affects long-run wine tourism intentions. Supported Note: n=553. Table 6. Correlations and descriptive statistics. AQWINT CPH WI ALWTINT LRWTINT WORSEINC SUPLOCW Acquired interest in wine during the lockdown (AQWINT) 3.5 (1.77) Covid-related fear and anxiety (CPH) 0.058 3.5 (1.63) Involvement with wine (WI) 0.662*** 0.058 5.2 (1.35) Wine tourism intentions after lockdown (ALWTINT) 0.404*** 0.004 0.494*** 5.1 (2.02) Future wine tourism intentions (LRWTINT) 0.466*** 0.102*** 0.640*** 0.624*** 5.9 (1.58) Worse income after Covid (WORSEINC) 0.109*** 0.106*** 0.149*** 0.171*** 0.131*** 0.3 (0.46) Willingness to support local wineries (SUPLOCW) 0.129*** 0.041 0.123*** 0.139*** 0.194*** 0.050 6.0 (1.35) Note: Mean (Std. Dev.) on the diagonal. *** p < .01 ** p < .05. 100 Giulia Gastaldello, Florine Livat, Luca Rossetto French sub-sample (CPH -> ALWTINT France β = .04; p = .48; WORSEINC -> ALWTINT France β = - .05; p= .86) but they are for the Italian one. Particularly, CPH has a signifi cant negative impact on ALWTINT (CPH -> ALWTINT Italy β = - .11; p < .001) while a worse income (WORSEINC) positively predicts short-term wine tourism intentions (WORSEINC -> ALWTINT Ita- ly β = .51; p < .001). Results of multigroup comparisons are summarized in Figure 3. Country-moderated mediation effects have been further explored. No signifi cant diff erences emerged for CPH mediation between the two groups (χ2 (352, 553) = 3.42, p = .18). Similarly, the mediation of AQWINT on the eff ect of WI on LRWTINT is not signifi cantly diff erent between France and Italy (χ2 (352, 553) = 3.80, p = .15). A signifi cant diff erence exists for the media- tion of AQWTINT on WI and ALWTINT (χ2 (352, 553) = 11.39, p = .003). Particularly, the indirect eff ect of WI on ALWTINT is positive for Italian respondents while it is negative for French wine tourists, despite poorly signifi cant (Italy β = .15; p < .004; France β = -.15; p =.092). Table 7. Mean and standard deviation of the variables included in the SEM by group. France (n=141) Italy (n=412) Mean St.Dev Mean St.Dev Involvement with wine (WI) 4.9 1.36 5.4 1.32 Acquired interest in wine during the lockdown (AQWINT) 3.0 1.79 3.6 1.73 Covid-related fear and anxiety (CPH) 3.4 1.46 3.7 1.54 Wine tourism intentions aft er lockdown (ALWTINT) 4.3 2.06 5.3 1.93 Future wine tourism intentions (LRWTINT) 4.7 1.85 6.3 1.25 Willingness to support local wineries (SUPLOCW) 6.1 1.24 5.9 1.39 Note: n=553; Italy n=412; France n=141. Figure 3. Multigroup comparisons between Italy and France. Note: n = 553; ***p < 0.01; **p < 0.05; *p < 0.1. Th e fi rst result refers to Italy, the second to France. Signifi cant results are reported in bold. 101Does Covid scare wine travelers? Evidence from France and Italy 5. DISCUSSION AND CONCLUSION The present study is among the first to provide a comprehensive overview on how an unprecedented event like the pandemic affected wine tourists’ behav- ioural intentions considering both positive and negative factors. To do so, we focus on two major wine tourism actors which have been severely hit by Covid-19: Italy and France. Generally, this analysis suggests the pandemic boosted wine tourism intentions rather than limiting them. Particularly, a greater share of wine tourists from both countries is willing to travel outside their region of residence after the lockdown, either to a different region or to another European country. Diversely, the share of tourists willing to travel to a neighbouring wine region is significantly smaller. Both the average length of stay in the wine region and the planned budget for a wine holi- day record an increase compared to pre-Covid, despite a considerable share of respondents declaring a worse eco- nomic situation following the pandemic. This observa- tion is consistent with the overnight stays peak recorded between July and August 2020 in both countries, when most Covid limitations were lifted. For the future wine tourism research agenda, it would be interesting to eval- uate whether the pandemic encouraged wine holidays instead of other trips among (wine) tourists. A switch from hotels to agriturismo emerged in the Italian sample, which does not appear to be connected to fear of contagion. National tourism statistics support this tendency since, compared to 2019, overnight stays in accommodations other than hotels (e.g., agriturismo, camping) recorded a lower decrease (-45%) than hotel ones (-56%) in 2020. Moreover, they grew more than hotel stays in 2021 (+27%, compared to +19% for hotels), and are therefore recovering faster from the 2020/2019 drop: while the 2021/2019 variation for hotels is still above -40%, other accommodations raised to -28%. Further research is needed to verify the extent of such behavioural changes and to explore their drivers. In our study, Covid-induced fear and anxiety (CPH) only shows a minor and poorly significant negative effect on wine tourism intentions after the lockdown (ALWTINT). This is despite the data collection time- frame, i.e. after the first wave of infection, when infor- mation on the virus and potential treatments was still scarce. Moreover, CPH does not mediate the relationship between future wine tourism intentions (LRWTINT) and intention to go on a post-lockdown wine holiday. The mild negative impact of CPH may be explained by the fact that wine tourists tend to be older than regular tourists, and the Covid-mortality rate is greater for the elderly [81]. Nevertheless, in line with existing studies [e.g., 35], CPH does not constitute a substantial deterrent to wine holidays. Although more research is required, we can reasonably connect this outcome to a higher per- ceived safety of rural destinations (like wine regions) compared to city ones [6]. This hypothesis is reinforced by recent findings showing how the threat of Covid intensifies consumers’ tendency to avoid crowding [82], which is easier in rural area. It should be noted that the impact of CPH is remarkably higher for the Italian sample, where its direct effect on wine tourism intentions after the lock- down is negative and significant (β - .11, p < .01). At the same time, it is non-significant for French respond- ents. Trust in official communications may have played a role in determining this country difference since, as Villacé-Molinero et al. [68] highlighted, they impact on the likelihood to stick to travel plans. Therefore, this is an essential factor to be considered by future research on the topic. The fact that AQWINT in lockdown signif i- cantly affects post-lockdown wine tourism intentions (ALWTINT) suggests that the proper communication strategy can help attracting wine tourists ahead of time. The prolonged duration of the Covid pandemic enhances the relevance of this finding, drawing attention on the strategic role played by virtual wine content and social- media marketing in reaching a wider audience and retain existing consumers during infection peaks. By fostering an increase of online content use, Covid has also boosted their long-term marketing potential in reducing the time and financial investment for wine tourists approaching unknown wineries and wine regions. The effect of such activities, though, may vary from country to country. Indeed, the influence of situational wine involvement (AQWINT) on post-lockdown wine tourism intentions (ALWTINT) shows a significant direct effect only for the Italian subsample (β .20; p < .001). The same variable is a also a mediator of personal involvement with wine (WI) on ALWTINT for both French and Ital- ians, while playing a greater and positive role for the lat- ter. Summing up, while in Italy situational involvement is an antecedent of short-term wine tourism intentions inde- pendently from involvement with wine, its effect is exclu- sively connected to the latter variable in France. Nevertheless, ss past studies suggest [33, 62], the sig- nificant mediation of AQWINT on the path from WI to wine tourism intentions supports the relevance of situa- tional involvement in enhancing the predictive power of WI. Academically, this finding paves the way to further research exploring the role of situational involvement in predicting wine tourism intentions and behaviour. 102 Giulia Gastaldello, Florine Livat, Luca Rossetto WI further confirmed to be a key antecedent of long-term wine tourism intentions, [52, 53, 54]. The remarkable standard deviation observed for WI high- lights the present sample includes wine tourists pos- sessing different degrees of interest and involvement with wine: a characteristic that may impact their future behavioural intentions. Future studies should address this issue and analyse group differences in wine tourism behaviour after the Covid outbreak based on respond- ents’ profiles as wine consumers, which is beyond the scope of this study. Solidarity, intended as the willingness to support local wineries by purchasing their products (SUPLOCW), emerged as a noteworthy driver of long-term wine tour- ism intentions. This finding is in line with proxim- ity being a key driver of wine tourism [25], which is also supported by the remarkable share of day-trippers in the sample. Moreover, it highlights the strong connection between the wine tourism phenomenon and support to rural communities through direct sales [66] and, on a greater scale, the vital role wine tourism can have as a form of sustainable tourism, answering rising concerns of tourism growth in the context of climate change [10]. Winery owners and tourism stakeholders should build on the willingness to support local businesses to attract trav- ellers outside major city destinations, designing sustain- able itineraries and experiences in rural areas. Post-lockdown wine tourism intentions (ALWTINT) seem to benefit of proximity as well, being positively impacted by negative repercussions of Covid-19 on house- hold income. So, in a sense, trips to close wine areas may represent an attractive and affordable getaway for families suffering the negative financial repercussions of Covid-19. This is true especially for the Italian subsample, where the effect is significant and not negligible (β 0.12; p < .01). Despite some researchers argue that the pandemic brought people attention on society problems [63], in our model solidarity with local winemakers after the Covid-19 crisis does not impact intentions to go on a wine holiday after the lockdown significantly. This out- come may be the result of risks connected to travelling representing a too high price to pay to prioritize collec- tive wellbeing, since the potential losses associated with Covid infection include health issues. Whilst offering a comprehensive overview on a still unexplored topic, the present study comes with some limitations, which are mostly connected to operational difficulties in collecting data. Notably, a relevant size difference between the two subpopulations exists. In this respect, data analysis relied on weighted estimates based on the French and the Italian group sizes. Some heterogeneity in terms of wine tourism intentions is also present between the two countries. The nature of such Country-based behavioural differences calls for further research, while the present study results represent an exploratory step forward to their comprehension. To conclude, the pandemic has deeply impact- ed tourism dynamics, inducing changes in travellers’ behaviour that call for fast, innovation-based respons- es [68]. Moreover, the emergence and re-emergence of lethal viruses have become increasingly frequent and worrying in the last decade, notably for the ease of transmission fostered by international travel [83]. Cov- id itself is still undefeated, and new viral variants are emerging. The findings of this study, therefore, provide wine tourism stakeholders with relevant information on how such unprecedented circumstances impact wine tourists’ behaviour and to effectively plan a recovery strategy accordingly. Academically, this research repre- sents important progress to wine tourism research as, differently from many past studies, it provides a compre- hensive view of behavioural intentions by simultaneously modelling positive and negative drivers of intentions: an improvement which is very much needed to avoid unde- sired myopias connected to the important role played by constraints in behavioural research [84]. REFERENCES [1] E. Calgaro, K. Lloyd, Sun, sea, sand and tsunami: examining disaster vulnerability in the tourism community of Khao Lak, Thailand, Singap. J. Trop. Geogr. 29(3) (2008) 288–306. 10.1111/j.1467- 9493.2008.00335.x. [2] M. Mazzocchi, A. Montini, Earthquake effects on tourism in central italy, Ann. Tour. Res. 28(4) (2021) 1031–1046. [3] B. Faulkner, S. Vikulov, Katherine, Washed out one day, back on track the next: a post-mortem of a tourism disaster, Tour. Manag. 22(4) (2001) 331– 344. 10.1016/S0261-5177(00)00069-8. [4] C. Bonham, C. Edmonds, J. Mak, The Impact of 9/11 and Other Terrible Global Events on Tourism in the United States and Hawaii, J. Travel Res. 45 (2006) 99–110. 10.1177/0047287506288812. [5] P. Gut, S. Jarrell, Silver Lining on a Dark Cloud: The Impact of 9/11 on a Regional Tour- ist Destination, J. Travel Res. 46 (2007) 147–153. 10.1177/0047287507299590. [6] G. Song, F. Khan, M. Yang, Probabilistic assess- ment of integrated safety and security related abnormal events: a case of chemical plants, Saf. Sci. 113 (2019) 115–125. 10.1016/J.SSCI.2018.11.004. 103Does Covid scare wine travelers? Evidence from France and Italy [7] O. Gergaud, F. Livat, H. Song, Terrorism and Wine Tourism: The Case of Museum Attendance, J. Wine Econ., 13(4) (2018) 375–383. 10.1017/jwe.2018.41. [8] A. Fleischer, S. Buccola, War, terror, the tourism market in Israel, Appl. Econ. 34(11) (2002) 1335– 1343. [9] H. Song, S. Lin, Impacts of the Financial and Eco- nomic Crisis on Tourism in Asia, J. Travel Res. 49(1) (2010) 16–30. doi: 10.1177/0047287509353190. [10] S. Gössling, D. Scott, C. M. Hall, Pandemics, tourism and global change: a rapid assessment of COVID-19. J. Sust. Tour. 29(1) (2020) 1-20. 10.1080/09669582.2020.1758708 [11] H.-I. Kuo, C.-L. Chang, B.-W. Huang, C.-C. Chen, M. McAleer, Estimating the Impact of Avi- an Flu on International Tourism Demand Using Panel Data, Tour. Econ. 15(3) (2009) 501–511. 10.5367/000000009789036611. [12] H. I. Kuo, C. C. Chen, W. C. Tseng, L. F. Ju, B. W. Huang, Assessing impacts of SARS and Avian Flu on international tourism demand to Asia, Tour. Manag. 29(5) (2008) 917–928. 10.1016/J.TOUR- MAN.2007.10.006. [13] M. McAleer, B. W. Huang, H. I. Kuo, C. C. Chen, C. L. Chang, An econometric analysis of SARS and Avian Flu on international tourist arrivals to Asia, Environ. Model. Softw. 25(1) (2010) 100–106. 10.1016/J.ENVSOFT.2009.07.015. [14] M. Novelli, L. Gussing Burgess, A. Jones, B. W. Ritchie, ‘No Ebola…still doomed’ – The Ebola- induced tourism crisis, Ann. Tour. Res. 70 (2018) 76–87. 10.1016/J.ANNALS.2018.03.006. [15] M. Gallivanid, B. Oppenheimid, N. K. Madhav, 2019. Using social media to estimate Zika’s impact on tourism: #babymoon, 2014-2017. PLoS One 14(2), e0212507. 10.1371/journal.pone.0212507. [16] J. Rossello, M. Santana-Gallego, W. Awan, Infec- tious disease risk and international tourism demand, Heal. Policy Plan. 32(4) (2017) 538–548. 10.1093/heapol/czw177. [17] J. Durocher, Recovery Marketing: What to Do after a Natural Disaster, Cornell Hotel Restaur. Adm. Q., 35(2) (1994) 66–70. 10.1177/001088049403500220. [18] H. H. Lean, R. Smyth, Asian Financial Crisis, Avi- an Flu and Terrorist Threats: Are Shocks to Malay- sian Tourist Arrivals Permanent or Transitory?, Asia Pacific J. Tour. Res. 14(3) (2009) 301–321. 10.1080/10941660903024034. [19] J. H. Huang, J. C. H. Min, Earthquake devastation and recovery in tourism: the Taiwan case, Tour. Manag. 23(2) (2002) 145–154. 10.1016/S0261- 5177(01)00051-6. [20] B. Rittichainuwat, Ghosts: A travel barrier to tour- ism recovery, Ann. Tour. Res. 38(2) (2011) 437– 459. 10.1016/J.ANNALS.2010.10.001. [21] A. Biran, W. Liu, G. Li, V. Eichhorn, Consuming post-disaster destinations: The case of Sichuan, China, Ann. Tour. Res. 47 (2014) 1–17. 10.1016/J. ANNALS.2014.03.004. [22] H. Song, R. T. R. Qiu, J. Park, A review of research on tourism demand forecasting: Launching the Annals of Tourism Research Curated Collection on tourism demand forecasting, Ann. Tour. Res. 75 (2019) 338–362. 10.1016/J.ANNALS.2018.12.001. [23] T. Dogru, U. Bulut, Is tourism an engine for eco- nomic recovery? Theory and empirical evidence, Tour. Manag. 67 (2018) 425–434. 10.1016/J.TOUR- MAN.2017.06.014. [24] L. Cheng, J. Zhang, Is tourism development a cata- lyst of economic recovery following natural disas- ter? An analysis of economic resilience and spatial variability, Curr. Issues Tour. 23(20) (2020) 2602– 2623. 10.1080/13683500.2019.1711029. [25] D. Getz,G. Brown, Critical success factors for wine tourism regions: A demand analysis, Tour. Manag. 27(1) (2006) 146–158. 10.1016/j.tour- man.2004.08.002. [26] V. Boatto, L. Galletto, L. Barisan, F. Bianchin, The development of wine tourism in the Conegliano Valdobbiadene area, Wine Econ. Policy. 2(2) (2013) 93–101. 10.1016/J.WEP.2013.11.003. [27] R. Garibaldi, Rapporto sul Turismo Enogastro- nomico Italiano 2020. Trend e tendenze. (Accessed 15 June 2021). https://www.robertagaribaldi.it/rap- porto-sul-turismo-enogastronomico/ [28] J. Winfree, C. McIntosh, T. Nadreau, An eco- nomic model of wineries and enotourism, Wine Econ. Policy. 7(2) (2018) 88–93. 10.1016/J. WEP.2018.06.001. [29] S. Castriota, M. Delmastro, The Economics of Col- lective Reputation: Evidence from the Wine Indus- try, Am. J. Agric. Econ. 97(2) (2015) 469–489. 10.1093/ajae/aau107. [30] B. Sparks, Planning a wine tourism vacation? Fac- tors that help to predict tourist behavioural inten- tions, Tour. Manag. 28(5) (2007) 1180–1192. 10.1016/j.tourman.2006.11.003. [31] D. Alonso, R. A. Fraser, D. A. Cohen, Explor- ing wine tourism in New Zealand: The visitors’ points of views, Tour. Anal. 13(2) (2008). 171–180. 10.3727/108354208785664247. [32] C. Afonso, G. M. Silva, H. M. Gonçalves, M. Duarte, The role of motivations and involvement in wine tourists’ intention to return: SEM and fsQCA 104 Giulia Gastaldello, Florine Livat, Luca Rossetto findings, J. Bus. Res. vol. 89 (2018). 10.1016/j.jbus- res.2017.11.042. [33] G. P. Brown, M. E. Havitz, D. Getz, Relationship between wine involvement and wine-related travel, J. Travel Tour. Mark., 21(1) (2007) 31–46. 10.1300/ J073v21n01_03. [34] S. Gammon, G. Ramshaw, Distancing from the Present: Nostalgia and Leisure in Lock- d ow n . L e i s . S c i . 4 3 ( 1 - 2 ) ( 2 0 2 0 ) 1 – 7 . 10.1080/01490400.2020.1773993. [35] J. M. Luo, C. F. Lam, Travel anxiety, risk attitude and travel intentions towards ‘travel bubble’ desti- nations in Hong Kong: Effect of the fear of COV- ID-19, Int. J. Environ. Res. 17(21) (2020) 1–11. 10.3390/ijerph17217859. [36] M. A. Mamun, M. D. Griffiths, First COVID-19 suicide case in Bangladesh due to fear of COV- ID-19 and xenophobia: Possible suicide preven- tion strategies, Asian J Psychiatr 51 (2020) 102073. 10.1016/j.ajp.2020.102073. [37] D. K. Ahorsu, C.-Y. Lin, V. Imani, M. Saffari, M. D. Griffiths, A. H. Pakpour, The Fear of COVID-19 Scale: Development and Initial Validation., Inter- national journal of mental health and addiction. (2020) 1–9. 10.1007/s11469-020-00270-8. [38] N. N. De Hoog, W. W. Stroebe, J. B. De Wit, The processing of fear-arousing communications: How biased processing leads to persuasion, Soc. Influ. 3(2) (2008) 84–113. [39] D. A. Clark, A. T. Beck, Cognitive therapy of anxi- ety disorders: Science and practice. Guilford Press, 2011. [40] H. Lazaratou, T. Paparrigopoulos, C. Anomitri, N. Alexandropoulou, G. Galanos, C. Papageorgiou, Sleep problems six-months after continuous earth- quake activity in a Greek island., Psychiatr. Psychi- atr. 29(1) (2018) 25-33. [41] J. M. Longman et al., Rationale and methods for a cross-sectional study of mental health and wellbeing following river flooding in rural Aus- tralia, using a community-academic partnership approach, BMC Public Health 19(1) (2019) 1–15. [42] W. Dai et al., Prevalence of acute stress disorder among road traffic accident survivors: a meta-anal- ysis., BMC psychiatry. 18(1) (2018) 1–11. [43] L. Yu Lin et al., The immediate impact of the 2019 novel coronavirus (COVID-19) outbreak on sub- jective sleep status, Sleep Med., 77 (2020) 348-354. 10.1016/j.sleep.2020.05.018. [44] I. Arpaci, K. Karataş, M. Baloğlu, The develop- ment and initial tests for the psychometric prop- erties of the COVID-19 Phobia Scale (C19P-S), Pers. Individ. Dif. 164 (2020), 110108. 10.1016/j. paid.2020.110108. [45] J. Schijven, L. C. Vermeulen, A. Swart, A. Mei- jer, E. Duizer, A. M. de Roda Husman 2020, Exposure assessment for airborne transmis- sion of SARS-CoV-2 via breathing, speaking, coughing and sneezing, medRxiv, 20144832. 10.1101/2020.07.02.20144832. [46] N. Michaelidou, S. Dibb, Consumer involvement: a new perspective, Mark. Rev. 8(1) (2008) 83–99. 10.1362/146934708x290403. [47] A. J. Broderick, R. D. Mueller, A Theoretical and Empirical Exegesis of the Consumer Involvement Construct: The Psychology of the Food Shop- per, J. Mark. Theory Pract. 7(4) (1999) 97–108. 10.1080/10696679.1999.11501855. [48] Prebensen, Nina K., et al., Motivation and involve- ment as antecedents of the perceived value of the destination experience. J. of Trav. Res. 52(2) (2013) 253-264. [49] J. L. Zaichkowsky, Measuring the involvement con- struct, J. Consum. Res. 12(12) (1985) 341‐52. [50] O. A. Ogbeide, J. Bruwer, Enduring involve- ment with wine: Predictive model and meas- urement, J. Wine Res. (3) (2013) 210–226. doi: 10.1080/09571264.2013.795483. [51] L. Lockshin, T. Spawton, Using Involvement and Brand Equity to Develop a Wine Tourism Strategy, J. Wine Mark. 13(1) (2001) 72–81. [52] J. Bruwer, I. Lesschaeve, Wine Tourists’ Destination Region Brand Image Perception and Antecedents: Conceptualization of a Winescape Framework, J. Travel Tour. Mark., 29(7) (2012) 611–628. doi: 10.1080/10548408.2012.719819. [53] J. Bruwer, C. Buller, Product involvement, brand loyalty and country-of-origin (COO) brand pref- erences of Japanese wine consumers, J. Wine Res. 24(1) (2013) 38–58. [54] D. Roe, J. Bruwer, Self-concept, product involve- ment and consumption occasions: Exploring fine wine consumer behaviour, Br. Food J. 119(6) (2017) 1362–1377. doi: 10.1108/BFJ-10-2016-0476. [55] I. Lesschaeve, J. Bruwer, The importance of con- sumer involvement and implications for new prod- uct development, in Consumer-Driven Innovation in Food and Personal Care Products, H. MacFie and S. R. Jaeger, Eds. Cambridge: Woodhead Pub- lishing Ltd, (2010) 386–423. [56] J. Bruwer, J. Huang, Wine product involvement and consumers’ BYOB behaviour in the South Australi- an on‐premise market., Asia Pacific J. Mark. Logist. 24(3) (2012) 461–481. 105Does Covid scare wine travelers? Evidence from France and Italy [57] G. Laurent, J. Kapferer, Measuring consumer involvement profiles, J. Mark. Research. 22(1) (1985) 41–53. [58] Gursoy, Dogan, and Erdogan Gavcar. International leisure tourists’ involvement profile. Ann. Tour. Res. 30(4) (2003) 906-926. [59] J. C. Hong, M. Y. Hwang, M. C. Liu, H. Y. Ho, Y. L. Chen, Using a ‘prediction-observation-explana- tion’ inquiry model to enhance student interest and intention to continue science learning predicted by their Internet cognitive failure, Comput. Educ., 72 (2014) 110–120. 10.1016/j.compedu.2013.10.004. [60] A. Nella, E. Christou, Segmenting Wine Tour- ists on the Basis of Involvement with Wine, J. Travel Tour. Mark. 31(7) (2014) 783–798. doi: 10.1080/10548408.2014.889639. [61] G. Easterbrook-Smith, By Bread Alone: Baking as Leisure, Performance, Sustenance, During the COVID-19 Crisis, Leis. Sci. 43(1–2) (2021) 36–42. 10.1080/01490400.2020.1773980. [62] M. Havitz, E. Mannel, C. Roger, Enduring involve- ment, situational involvement, and flow in leisure and non-leisure activities. J. Leis. Res. 37(2) (2005) 152-177 [63] A. W. Cappelen et al. Solidarity and fairness in times of crisis, J. Econ. Behav. Org. 186 (2021) 1-11. [64] C. Mauracher, I. Procidano, G. Sacchi, Wine tour- ism quality perception and customer satisfaction reliability: the Italian Prosecco District. J. Wine Res. 27(4) (2016) 284-299. [65] UNWTO, Georgia Declaration on Wine Tourism. Fostering sustainable tourism development through intangible cultural heritage, (2016). [66] E. Giampietri, D. Koemle, X., Yu, A. Finco, Con- sumers’ sense of farmers’ markets: tasting sustain- ability or just purchasing food?. Sust. 8(11) (2016) 1157. https://doi.org/10.3390/su8111157 [67] A. D., Alonso, A. Bressan, M. O’Shea, V., Krajsic, Perceived benefits and challenges to wine tour- ism involvement: An international perspective. Int. J. Tour. Res. 17(1) (2015) 66-81. https://doi. org/10.1002/jtr.1967 [68] T. Villacé-Molinero, J. J. Fernández-Muñoz, A. Orea-Giner, L. Fuentes-Moraleda, Understand- ing the new post-COVID-19 risk scenario: Out- looks and challenges for a new era of tourism, Tour. Manag. 86 (2021) 104324. doi: https://doi. org/10.1016/j.tourman.2021.104324. [69] S. Park, B. Stangl, Augmented reality experiences and sensation seeking, Tour. Manag. 77 (2020) 104023. doi: 10.1016/J.TOURMAN.2019.104023. [70] F. Baltar, I. Brunet, Social research 2.0: vir- tual snowball sampling method using Face- book, Internet Res. 22(1) (2012) 57–74. doi: 10.1108/10662241211199960. [71] K. Sabin, L. G. Johnston, K. Sabin, Sampling hard- to-reach populations with respondent driven sam- pling, Methodol. Innov. Online. 5(2) (2010) 38–48. doi: 10.4256/mio.2010.0017. [72] G. Robins Sadler, H.-C. Lee, R. S.-H. Lim, J. Fuller- ton, Recruitment of hard-to-reach population sub- groups via adaptations of the snowball sampling strategy, Nurs. Heal. Sci. 12 (2010) 369–374. doi: 10.1111/j.1442-2018.2010.00541.x. [73] S. Charters, J. Ali-Knight, Who is the wine tourist?, Tour. Manag. 23(3) (2002) 311–319. doi: 10.1016/ S0261-5177(01)00079-6. [74] V. Asero, S. Patti, Wine tourism experience and consumer behavior: The case of sicily, Tour. Anal. 16(4) (2011) 431–442. doi: 10.3727/108354211X13 149079788936. [75] M. G. Brandano, L. Osti, M. Pulina, How motiva- tions and satisfaction influence wine tourists’ loy- alty? An analysis of the Italian case, Int. J. Cult. Tour. Hosp. Res. 13(1) (2018) 55–69. doi: 10.1108/ IJCTHR-04-2018-0054. [76] J. Fountain, S. Charters, L. Cogan-Marie. The real Burgundy : negotiating wine tour- ism, relational place and the global country- side. Tour. Geogr. 23(5-6) (2021) 1116-1136. 10.1080/14616688.2020.1713880. [77] A. Zatori, M. K. Smith, L. Puczko. Experience- involvement, memorability and authenticity: The service provider’s effect on tourist experience. Tour. Manag. 67 (2018) 111-126. 10.1016/j.tour- man.2017.12.013 [78] A. B. Costello, J. W. Osborne, Best practices in exploratory factor analysis: four recommendations for getting the most from your analysis, Pract. Assessment, Res. Eval. (10)7 (2005) 1–9. [79] J. F. Hair, W. C. Black, B. J. Babin, R. E. Anderson, Multivariate Data Analysis. U.K.: Cenage, 2018. [80] E. Ryu, J. Cheong, Comparing indirect effects in different groups in single-group and multi-group structural equation models, Front. Psychol. 8 (2017) 747. doi: 10.3389/fpsyg.2017.00747. [81] S. S. Bhopal, R. Bhopal, Sex differential in COV- ID-19 mortality varies markedly by age, Lancet. 396(10250) (2020) 532–533. doi: 10.1016/S0140- 6736(20)31748-7. [82] I.-J. Park, J. Kim, S. (Sam) Kim, J. C. Lee, M. Gir- oux, Impact of the COVID-19 pandemic on trave- lers’ preference for crowded versus non-crowd- 106 Giulia Gastaldello, Florine Livat, Luca Rossetto ed options, Tour. Manag. (2021) 104398. doi: 10.1016/J.TOURMAN.2021.104398. [83] F. Houghton, Geography, global pandemics & air travel: Faster, fuller, further & more frequent, J. Infect. Public Health. 12(3) (2019) 448–449. doi: 10.1016/j.jiph.2019.02.020. [84] M. Cho, M. A. Bonn, R. A. Brymer. A constraint- based approach to wine tourism market segmenta- tion. J. Hosp. Tour. Res. 41(4) (2017) 415-444. Doi https://doi.org/10.1177/1096348014538049