Journal of Hospitality and Tourism Management

Journal of Hospitality and Tourism Management 57 (2023) 225–235 230 6.3. Model comparisons We proceeded with comparing four competing models (Fig. 3) as per Anderson and Gerbing’s (1988) approach. Upon examining the results in Table 6, it was evident that some model fit indices for both Model 1 and Model 2 did not meet the recommended criteria, signifying that these models were not deemed acceptable. Conversely, Model 4 demonstrated satisfactory fit indices across all criteria, affirming the viability of conceptualizing HDC as a second-order construct with a four-factor structure. 6.4. Cross validity To assess the cross validity of the HDC scale, we employed invariance analysis as suggested by Xie et al. (2022). The Study 3 sample was divided into two sub-samples using two distinct approaches: gender-based grouping and random splitting. The invariance test results Table 1 Participants’ demographic information. Category Study2 Study3 Category Study2 Study3 Gender Male 39.7% 36.3% Marital status Married 37.3% 54.2% Female 60.3% 63.7% Unmarried 62.7% 45.8% Education Junior high school or below 9.8% 7.7% Monthly income (CNY) 2500≤ 32.4% 17.1% Senior high school 20.6% 13.2% 2501–5000 40.2% 31.6% Junior college 25.5% 21.6% 5001–10000 19.6% 36.0% Bachelor’s degree 37.7% 54.0% 10001-20000 4.4% 12.2% Master’s degree or above 6.4% 3.5% ≥20001 3.4% 3.1% Age 20 or below 10.8% 6.5% Department Front office 13.7% 11.8% 20~29 47.1% 54.6% Food and beverage 43.6% 24.6% 30~39 18.6% 27.3% Housekeeping 8.3% 15.1% 40~49 17.6% 9.6% Entertainment 1.5% 0.2% 50 or above 5.9% 2.0% Finance 3.9% 6.5% Work experience ≤1 year 47.1% 23.2% Human resources 4.4% 8.1% 1<,≤3 years 8.3% 20.8% Sales 3.4% 6.1% 3<,≤5 years 9.3% 15.9% Security 2.0% 3.5% 5<,≤10 years 13.7% 30.1% Engineering 3.9% 2.2% >10 years 21.6% 10.0% Information 2.0% 10.0% Position Junior staff 67.6% 50.5% Others 13.2% 11.8% Foreman 6.4% 15.1% Type Three-star 5.90% 13.00% Supervisor 10.8% 17.3% four-star 32.80% 35.00% Manager 10.8% 14.1% Five star 41.70% 36.30% Director 4.4% 3.0% other 19.60% 15.70% Table 2 Results of exploratory factor analysis. Construct Items Communality Loading Cronbach’s α Basic capacity BC2 0.50 0.65 0.87 BC3 0.55 0.73 BC4 0.73 0.87 BC5 0.59 0.80 BC7 0.67 0.59 Integration capacity IC1 0.63 0.58 0.86 IC2 0.59 0.60 IC5 0.60 0.83 IC6 0.72 0.75 Application capacity AC1 0.62 0.61 0.88 AC2 0.57 0.65 AC3 0.55 0.63 AC4 0.65 0.73 AC6 0.64 0.72 Optimization capacity OC1 0.65 0.68 0.93 OC2 0.73 0.80 OC3 0.72 0.78 OC4 0.75 0.84 OC5 0.70 0.88 OC6 0.61 0.72 Table 3 Results of confirmatory factor analysis. Dimensions Items Mean Loading CR AVE Cronbach’s α Basic capacity BC2 5.85 0.61 0.85 0.53 0.85 BC3 6.00 0.66 BC4 5.71 0.74 BC5 5.73 0.79 BC7 5.79 0.81 Integration capacity IC1 5.73 0.74 0.81 0.52 0.81 IC2 5.54 0.72 IC5 5.43 0.65 IC6 5.74 0.76 Application capacity AC1 5.61 0.65 0.81 0.46 0.81 AC2 5.47 0.68 AC3 6.02 0.62 AC4 5.69 0.69 AC6 5.50 0.74 Optimization capacity OC1 5.73 0.76 0.82 0.53 0.82 OC2 5.73 0.72 OC3 5.70 0.73 OC4 5.63 0.71 Table 4 Constructs’ correlations and the squared root of AVE. Dimensions Mean S.D. 1 2 3 4 1 Basic capacity 5.81 0.94 (0.73) 2 Integration capacity 5.60 0.95 0.63** (0.73) 3 Application capacity 5.66 0.95 0.50** 0.56** (0.68) 4 Optimization capacity 3.80 0.61 0.55** 0.63** 0.66** (0.73) Table 5 Discriminant validity—Heterotrait - Monotrait ratio (HTMT). Dimensions 1 2 3 4 1 Basic capacity 2 Integration capacity 0.77 3 Application capacity 0.61 0.69 4 Optimization capacity 0.66 0.77 0.82 L. Fan et al.

Journal of Hospitality and Tourism Management 57 (2023) 225–235 231 (Table 7) indicated that in both grouping methods, the model fit indices were satisfactory, and no significant discrepancies were observed (Gender split: △χ2 = 13.52, △df = 14, p = 0.49; Random split: △χ2 = 19.24, △df = 14, p = 0.16) between the groups. Moreover, the absolute differences in fit indices (i.e., IFI, TLI, CFI, GFI) were all below 0.01. These outcomes affirm the cross-sample validity of HDC. 6.5. Predictive validity To gauge HDC’s predictive validity, we utilized organizational resilience as the criterion variable. This choice was motivated by prior research that has investigated the link between digital capability and organizational resilience (He et al., 2023; Williams et al., 2017). Items drawn from the work of Melian-Alzola ´ et al. (2020) were employed to assess organizational resilience. These items were integrated into the Study 3 questionnaire for the purpose of testing predictive validity. The Fig. 3. Four competing models. L. Fan et al.

Journal of Hospitality and Tourism Management 57 (2023) 225–235 232 findings revealed a positive impact of HDC on organizational resilience (β = 0.63, p < 0.001). Moreover, basic capability (β = 0.44, p < 0.001), integration capability (β = 0.52, p < 0.001), application capability (β = 0.55, p < 0.001), and optimization capability (β = 0.58, p < 0.001) all exhibited positive correlations with organizational resilience. These outcomes underscore the robust predictive validity of the developed HDC scale. 7. Conclusions and discussion 7.1. Conclusions This study aims to develop a measurement scale for Hotel Digital Capability (HDC). Study 1 involved in-depth interviews to establish the dimensional structure of HDC. Study 2 and Study 3 focused on the development and validation of the HDC measurement scale. The findings consistently supported a four-factor structure for HDC and demonstrated the strong validity of the developed items. Digital technologies have fundamentally reshaped how hotels operate and manage their businesses. While previous research has explored the positive impacts of various digital technologies on hotel development, it often leaned towards the technical aspects of implementation (Zhang et al., 2022; Kao & Huang, 2023). However, it’s imperative to recognize that digital technology’s application in hotels encompasses a broader capability performance, integrating organizational practices and emphasizing long-term value creation. This digital capability empowers hotels to establish a customer-centric ecosystem. Through data collection and understanding of user behavior, hotels can create precise customer and digital consumption profiles. This optimization leads to enhanced customer experiences, greater loyalty, an expanded service portfolio, and heightened operational efficiency. Despite its significant impact, the concept of HDC has received limited attention, resulting in a dearth of reliable measurement tools. The HDC scale developed and validated in this study bridges this gap. It not only evaluates the overall level of digitalization within hotels but also forms a foundation for exploring the relationship between HDC and other critical constructs like organizational resilience. This scale, consequently, provides a robust measurement tool for quantitative research in related fields. 7.2. Theoretical implications Our study offers three significant contributions to the existing literature. First, it elucidates the four-dimensional structure of HDC, adding depth and breadth to the theoretical discourse on digital technology within the hotel industry. While previous research has explored the antecedents and outcomes of digital technologies in hotels (Hong et al., 2021; Khanra et al., 2021), much of it has focused on specific types of technologies (Chen et al., 2016; Park et al., 2021). These studies emphasize the importance of integrating digital technology in the hospitality sector, aiding hotel managers and proprietors in informed decision-making and strategic technology investments (Iranmanesh et al., 2022). In the context of the digital economy, the adoption of digital technology in hotels has become imperative. Beyond mere adoption, hotel managers must concentrate on cultivating digital capabilities rooted in digital technology. These capabilities play a pivotal role in harmonizing hotel operations, embedding managerial functions, and shaping strategic blueprints (Law & Jogaratnam, 2005). Despite attention in fields like manufacturing and the Internet, the dimensional structure of Hotel Digital Capability has been overlooked. Our research delineated the dimensional framework of HDC and thus filled this research void. Second, this study pioneers a hierarchical evolutionary model of digital capability from a dynamic standpoint, providing a fresh theoretical vantage point for HDC research. Prior research has explored digital capabilities from various angles (Annarelli et al., 2021; Lenka et al., 2017; Ritter & Pedersen, 2020). However, these studies often perceive digital capability as a static concept, failing to fully acknowledge its dynamic nature and evolution across dimensions. Our study interprets hotel digital capability through a hierarchical model of dynamic capabilities (Winter, 2003). This hierarchical perspective illuminates the interactions among the inherent dimensions of HDC. Third, this study forges a measurement scale for HDC through a multi-stage empirical approach. This scale furnishes a reliable tool for quantitative research in related domains and bolsters the progress of theoretical and empirical research on HDC. In the hospitality industry, digital capabilities predominantly revolve around customer service, encompassing the identification of customer needs, response to service requests, and post-service product lifecycle tracking. Consequently, the distinctive nature of the hospitality industry hampers the direct application of measurement items from other fields. In light of this, it is crucial to develop HDC that provides an accurate understanding of how the hospitality industry leverages digital technology. 7.3. Practical implications Our study offers practical insights. With the escalating importance of digital technology in the hotel industry, managers must prioritize HDC Table 6 Model comparisons of HDC. Goodness-fit-indices Model 1 Model 2 Model 3 Model 4 RMSEA (<.08) 0.10 0.12 0.04 0.05 SRMR (<.08) 0.07 0.31 0.03 0.04 CFI (>.9) 0.82 0.75 0.97 0.96 NFI (>.9) 0.80 0.72 0.94 0.93 TLI (>.9) 0.80 0.71 0.96 0.95 IFI (>.9) 0.82 0.75 0.97 0.96 GFI (>.8) 0.80 0.78 0.95 0.94 AGFI (>.8) 0.74 0.71 0.93 0.92 PNFI (>.5) 0.70 0.64 0.79 0.80 x2 816.62 1118.29 245.29 281.89 df 135 135 129 131 x2 /df (1 <, <5) 6.05 8.28 1.90 2.15 Note: Model 1 = First-order factor model; Model 2 = Four unrelated first-order factor model; Model 3 = Four correlated first-order factor model; Model 4 = Second-order four-factor model. Table 7 Cross validity. Goodness-fit-indices Gender split (n1 = 178 Male, n2 = 313 Female) Random split (n1 = 246, n2 = 245) Unconstrained model Measurement model Unconstrained model Measurement model RMSEA 0.04 0.04 0.04 0.04 IFI 0.95 0.95 0.96 0.96 TLI 0.94 0.94 0.95 0.95 CFI 0.95 0.95 0.96 0.96 GFI 0.88 0.88 0.91 0.91 ECVI 1.28 1.25 1.22 1.21 x2 459.83 473.35 430.18 449.43 df 258.00 272.00 258.00 272.00 L. Fan et al.

Journal of Hospitality and Tourism Management 57 (2023) 225–235 233 development. They can formulate strategic approaches by leveraging HDC’s dimensional structure to systematically foster digital capabilities. For basic capabilities, managers should align technology adoption with the hotel’s business trajectory, strategic objectives, and customer demands. Additionally, consideration should be given to the hotel’s type and positioning when configuring or upgrading management systems and hardware like front desk systems, central reservations, customer management, food and beverage, and revenue management systems. This approach significantly boosts the hotel’s digital infrastructure capabilities. In terms of integration capability, managers need to weave digital technology into both internal and external hotel processes. This integration expands data sources and streamlines data accumulation, thereby supporting future product development and decision-making. Focusing on application capability, managers should prioritize enhancing the digital prowess of back-end processes, including cybersecurity and the impact of digital technology on employee well-being. Coordinating front-end services and back-end systems through platforms or systems further amplifies service efficiency, efficacy, and productivity. Regarding optimization capability, digital technology empowers hotels to forge connections with both customers and employees. Managers should streamline operations by minimizing redundant tasks and nurturing a sense of enthusiasm among the workforce. Concurrently, hotels can provide customers with convenient, personalized services through online platforms, transforming them from passive recipients to real-time co-creators in a technologically enhanced experiential environment. 7.4. Limitations and future research Our study has a few limitations. First, the study focused on mid-to high-end hotels in specific regions of China, employing convenience sampling. Future research could broaden the geographical scope to compare digital capabilities across diverse contexts. Additionally, in the sample structure of Studies 2 and 3, the proportion of F&B department staff is relatively high, potentially impacting scale reliability and validity. Future studies could utilize stratified or cluster sampling to refine and compare the developed HDC scale. Second, HDC is subject to change with evolving digital technology. Future studies should reevaluate measurement item relevance and refine the scale with new items aligning with technological advancements and novel applications in the hotel industry. Funding statement Supported by the National Social Science Foundation of China (Grant No. 23CGL030). Declaration of generative AI and AI-assisted technologies in the writing process During the preparation of this work the authors used ChatGPT in order to improve the clarity of the manuscript. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication. Appendices. Table A.1 Basic information of interview participants Number Position Department Educational Type of business Work experience 1 General Manager / Bachelor’s degree International hotel More than 15years 2 General Manager / Bachelor’s degree International hotel More than 15years 3 General Manager / Bachelor’s degree local hotel More than 15years 4 General Manager / Bachelor’s degree local hotel More than 20years 5 General Manager / Junior college local hotel More than 15years 6 Financial Director Finance Bachelor’s degree local hotel More than 10years 7 Security Manager Security Technical secondary school local hotel More than 15years 8 Engineering Manager Engineering Junior college local hotel More than 15years 9 Revenue Director Sales Bachelor’s degree local hotel More than 10years 10 Marketing Director Sales Bachelor’s degree local hotel More than 20years 11 Catering Director Food and beverage Junior college local hotel More than 15years 12 Room manager Housekeeping Bachelor’s degree local hotel More than 20years 13 Front Office Director Front office Junior college local hotel More than 20years 14 HR Director Human resources Bachelor’s degree local hotel More than 20years 15 HR Director Human resources Bachelor’s degree International hotel More than 10years 16 CEO/Professor / Master’s degree International hotel More than 15years 17 Operations Manager / Bachelor’s degree International hotel More than 5years 18 Professor / Master’s degree college More than 15years 19 Lecturer / Master’s degree college More than 5 years Table A.2 Examples of Open Coding Results Interviewee (A) Original Interview Data Open coding Initial concept …. … Because I am in the marketing department, I use examples from marketing. We developed a WTP system. What is it? This system is based on Mini Program. For example, consider a scenario where a salesman goes out to visit customers with his cell phone after a meeting in the morning. When he visits every customer, he opens WTP, enters the customer’s name, position, telephone number, the content of the visit, and then clicks OK in the system. After confirmation, the system will capture the time and place, and a complete visit process will end. And so on. After the salesman completes his work in that day and returns to the hotel, a report or a map is generated in the hotel’s backend information system. This map can be visualized, and records the time points, where he went, and what he said to his customers. Thus, the whole work track is recorded. By borrowing this report, I can clearly understand the job A-1 marketing management system A-2 online marketing tasks for employees A-3 digital management of offline marketing A-4 use digital technology to improve marketing efficiency A-5 realize precise marketing based upon data analytics (continued on next page) L. Fan et al.

Journal of Hospitality and Tourism Management 57 (2023) 225–235 234 Table A.2 (continued ) Interviewee (A) Original Interview Data Open coding Initial concept route of the salesman, and use it to evaluate whether the route is reasonable and efficient. After a certain time period, when enough data has been accumulated, the output of each customer in the hotel fluctuates. For example, from January to December, if the purchase fluctuates, no matter whether it increases or decreases, in general, the salesman’s visit to the customer should correspond to its fluctuation. If the purchase drops, his visit record should provide the reasons. What are the related reasons? If the issues are not resolved, the salesman’s visit can be unprofessional. This is digital management supporting offline marketing. This system is already in use, and it’s basically such an idea …. … A-6 hotel marketing business data integration A-7 supervise employees’ work with digital technology A-8 support managers’ decision with data analytics Appendix B. Measurement items of HDC Table B.1 The initial list of items Dimensions Items Scale Items Basic Capability BC1 The hotel has an automatic fire alarm system. BC2 The hotel has an intelligent card system. BC3 The hotel has an intelligent lighting control system. BC4 The hotel has an intelligent room control system. BC5 The hotel is equipped with a multimedia information inquiry system. BC6 The hotel is equipped with intelligent service robots. BC7 On the whole, the hotel has good digital equipment and platforms. Integration Capability IC1 Hotel information system supports the integration of workflow among departments. IC2 Hotel information system supports data sharing among departments. IC3 Hotel information system supports data sharing with cooperative enterprises. IC4 Hotel information system can collect different types of customer information. IC5 Hotel information system can collect multi-channel market information. IC6 Hotel information system supports information communication among various departments. Application Capability AC1 The hotel uses WeChat platform to carry out customer services. AC2 The hotel uses live-broadcasting platform/short video to carry out marketing promotion. AC3 The hotel uses e-commerce booking platform (e.g., Ctrip/Fliggy) to attract customers. AC4 The hotel has developed online shopping malls and other sales platforms. AC5 The hotel provides self-check-in service to customers. AC6 Online customer service system of the hotel can quickly respond to customer instructions. AC7 Hotel service personnel can submit service information online (e.g., online report for maintenance). Optimization Capability OC1 The hotel can use digital services to improve customer satisfaction OC2 The hotel can use digital services to shape their corporate image. OC3 The hotel s can use the information system to support the staff work. OC4 The hotel can use big data to help managers make decisions. OC5 The hotel can use information systems to improve staff productivity. OC6 The hotel will bring digital development into its future development goals. Table B.2 The final list of items Dimensions Items Scale Items Basic Capability BC1 The hotel has an automatic fire alarm system. BC2 The hotel has an intelligent card system. BC3 The hotel has an intelligent lighting control system. BC4 The hotel has an intelligent room control system. BC5 The hotel is equipped with a multimedia information inquiry system. BC7 On the whole, the hotel has good digital equipment and platforms. Integration capability IC1 Hotel information system supports the integration of workflow among departments. IC2 Hotel information system supports data sharing among departments. IC3 Hotel information system supports data sharing with cooperative enterprises. IC4 Hotel information system can collect different types of customer information. IC5 Hotel information system can collect multi-channel market information. IC6 Hotel information system supports information communication among various departments. Application Capability AC1 The hotel uses WeChat platform to carry out customer services. AC2 The hotel uses live-broadcasting platform/short video to carry out marketing promotion. AC3 The hotel uses e-commerce booking platform (e.g., Ctrip/Fliggy) to attract customers. AC4 The hotel has developed online shopping malls and other sales platforms. AC6 Hotels use online systems (e.g., WeChat group and online reviews) to manage customer relationships. AC7 Hotels use information system to manage business information (e.g., updating room status) Optimization Capability OC1 The hotel can use digital services to improve customer satisfaction OC2 The hotel can use digital services to shape their corporate image. OC3 The hotel s can use the information system to support the staff work. OC4 The hotel can use big data to help managers make decisions. OC5 The hotel can use information systems to improve staff productivity. OC6 The hotel will bring digital development into its future development goals. L. Fan et al.

Journal of Hospitality and Tourism Management 57 (2023) 225–235 235 References Anderson, J. C., & Gerbing, D. W. (1988). Structural equation modeling in practice: A review and recommended two-step approach. Psychological Bulletin, 103(3), 411–423. Annarelli, A., Battistella, C., Nonino, F., Parida, V., & Pessot, E. (2021). Literature review on digitalization capabilities: Co-Citation analysis of antecedents, conceptualization and consequences. Technological Forecasting and Social Change, 166, Article 120635. Braun, T., & Sydow, J. (2019). Selecting organizational partners for interorganizational projects: The dual but limited role of digital capabilities in the construction industry. Project Management Journal, 50(4), 398–408. Busulwa, R., Pickering, M., & Mao, I. (2022). Digital transformation and hospitality management competencies: Toward an integrative framework. International Journal of Hospitality Management, 102, Article 103132. Chen, M. M., Murphy, H. C., & Knecht, S. (2016). An importance performance analysis of smartphone applications for hotel chains. Journal of Hospitality and Tourism Management, 29, 69–79. Copeland, J. R. M., Kelleher, M. J., Kellett, J. M., Gourlay, A. J., Gurland, B. J., Fleiss, J. L., & Sharpe, L. (1976). A semi-structured clinical interview for the assessment of diagnosis and mental state in the elderly: The geriatric mental state schedule: I. Development and reliability. Psychological Medicine, 6(3), 439–449. Coreynen, W., Vanderstraeten, J., van Witteloostuijn, A., Cannaerts, N., Loots, E., & Slabbinck, H. (2020). What drives product-service integration? An abductive study of decision-makers’ motives and value strategies. Journal of Business Research, 117, 189–200. Fokkema, M., & Greiff, S. (2017). How performing PCA and CFA on the same data equals trouble. European Journal of Psychological Assessment, 33(6), 399–402. Gibbs, C., MacDonald, F., & MacKay, K. (2015). Social media usage in hotel human resources: Recruitment, hiring and communication. International Journal of Contemporary Hospitality Management, 27(2), 170–184. Gong, Y., Yao, Y., & Zan, A. (2023). The too-much-of-a-good-thing effect of digitalization capability on radical innovation: The role of knowledge accumulation and knowledge integration capability. Journal of Knowledge Management, 27(6), 1680–1701. Gottschalk, M., Kuntz, J. C., & Prayag, G. (2022). TouRes: Scale development and validation of a tourist resilience scale. Tourism Management Perspectives, 44, Article 101025. Grace, D., Ross, M., & King, C. (2020). Brand fidelity: Scale development and validation. Journal of Retailing and Consumer Services, 52, Article 101908. Hair, J. F., Anderson, R. E., Tatham, R. L., & Black, W. C. (2010). Multivariate data analysis: A global perspective 7 pearson. Upper Saddle River: NJ. Han, S. H., Lee, J., Edvardsson, B., & Verma, R. (2021). Mobile technology adoption among hotels: Managerial issues and opportunities. Tourism Management Perspectives, 38, Article 100811. He, Z., Huang, H., Choi, H., & Bilgihan, A. (2023). Building organizational resilience with digital transformation. Journal of Service Management, 34(1), 147–171. He, X. G., Liang, Q. X., & Wang, S. L. (2019). Information technology, labor force structure and enterprise productivity -solving the mystery of “paradox of information technology productivity”. Management World, 35(9), 65–80. Henseler, J., Ringle, C. M., & Sarstedt, M. (2015). A new criterion for assessing discriminant validity in variance-based structural equation modeling. Journal of the Academy of Marketing Science, 43, 115–135. Heredia, J., Castillo-Vergara, M., Geldes, C., Gamarra, F. M. C., Flores, A., & Heredia, W. (2022). How do digital capabilities affect firm performance? The mediating role of technological capabilities in the “new normal”. Journal of Innovation & Knowledge, 7 (2), Article 100171. Hong, C., Choi, H. H., Choi, E. K. C., & Joung, H. W. D. (2021). Factors affecting customer intention to use online food delivery services before and during the COVID-19 pandemic. Journal of Hospitality and Tourism Management, 48, 509–518. Hou, G. W., & Liu, Q. Q. (2022). Network power and innovation performance: From the perspective of firm digital capability. Studies in Science of Science, 40(6), 1143–1152. Hua, N. (2020). Do information technology (IT) capabilities affect hotel competitiveness? Journal of Hospitality and Tourism Technology, 11(3), 447–460. Iranmanesh, M., Ghobakhloo, M., Nilashi, M., Tseng, M. L., Yadegaridehkordi, E., & Leung, N. (2022). Applications of disruptive digital technologies in hotel industry: A systematic review. International Journal of Hospitality Management, 107, Article 103304. Jeong, M., Lee, M., & Nagesvaran, B. (2016). Employees’ use of mobile devices and their perceived outcomes in the workplace: A case of luxury hotel. International Journal of Hospitality Management, 57, 40–51. Junior, J. C. D. S. F., & Maçada, A. C. G. (2020). Examiming digital capabilities and their role in the digital business performance. Revista Economia & Gestao, ˜ 20(56), 148–171. Kao, W. K., & Huang, Y. S. S. (2023). Service robots in full-and limited-service restaurants: Extending technology acceptance model. Journal of Hospitality and Tourism Management, 54, 10–21. Khanra, S., Dhir, A., Kaur, P., & Joseph, R. P. (2021). Factors influencing the adoption postponement of mobile payment services in the hospitality sector during a pandemic. Journal of Hospitality and Tourism Management, 46, 26–39. Khin, S., & Ho, T. C. (2018). Digital technology, digital capability and organizational performance: A mediating role of digital innovation. International Journal of Innovation Science, 11(2), 177–195. Kim, W. G., McGinley, S., Choi, H. M., Luberto, E., & Li, J. J. (2020). How does room rate and rate dispersion in US hotels fluctuate? Journal of Hospitality and Tourism Management, 44, 227–237. Kim, T., Suh, Y. K., Lee, G., & Choi, B. G. (2010). Modelling roles of task-technology fit and self-efficacy in hotel employees’ usage behaviours of hotel information systems. International Journal of Tourism Research, 12(6), 709–725. Kline, R. B. (2011). Principles and practice of structural equation modeling (3. Baskı). New York, NY: Guilford. Lamest, M., & Brady, M. (2019). Data-focused managerial challenges within the hotel sector. Tourism Review, 74(1), 104–115. Lau, A. (2020). New technologies used in COVID-19 for business survival: Insights from the Hotel Sector in China. Information Technology & Tourism, 22(4), 497–504. Law, R., & Jogaratnam, G. (2005). A study of hotel information technology applications. International Journal of Contemporary Hospitality Management, 17(2), 170–180. Lenka, S., Parida, V., & Wincent, J. (2017). Digitalization capabilities as enablers of value co-creation in servitizing firms. Psychology and Marketing, 34(1), 92–100. Liu, C., & Yang, J. (2021). How hotels adjust technology-based strategy to respond to COVID-19 and gain competitive productivity (CP): Strategic management process and dynamic capabilities. International Journal of Contemporary Hospitality Management, 33(9), 2907–2931. Li, L., Zhu, W., Wei, L., & Yang, S. (2022). How can digital collaboration capability boost service innovation? Evidence from the information technology industry. Technological Forecasting and Social Change, 182, Article 121830. MacKenzie, S. B., Podsakoff, P. M., & Podsakoff, N. P. (2011). Construct measurement and validation procedures in MIS and behavioral research: Integrating new and existing techniques. MIS Quarterly, 293–334. Mandal, S., & Dubey, R. K. (2020). Role of tourism IT adoption and risk management orientation on tourism agility and resilience: Impact on sustainable tourism supply chain performance. International Journal of Tourism Research, 22(6), 800–813. Melian-Alzola, ´ L., Fern´ andez-Monroy, M., & Hidalgo-Penate, ˜ M. (2020). Hotels in contexts of uncertainty: Measuring organisational resilience. Tourism Management Perspectives, 36, Article 100747. Melian-Gonz ´ alez, ´ S., & Bulchand-Gidumal, J. (2016). A model that connects information technology and hotel performance. Tourism Management, 53, 30–37. Morosan, C., & DeFranco, A. (2019). Mapping the impact of hotel promotional factors on consumers’ actual use of interactive systems in hotels. Journal of Hospitality and Tourism Technology, 10(2), 169–189. Park, S., Lehto, X., & Lehto, M. (2021). Self-service technology kiosk design for restaurants: An QFD application. International Journal of Hospitality Management, 92, Article 102757. Piccoli, G., Lui, T. W., & Grün, B. (2017). The impact of IT-enabled customer service systems on service personalization, customer service perceptions, and hotel performance. Tourism Management, 59, 349–362. Ritter, T., & Pedersen, C. L. (2020). Digitization capability and the digitalization of business models in business-to-business firms: Past, present, and future. Industrial Marketing Management, 86, 180–190. Sahadev, S., & Islam, N. (2005). Why hotels adopt ICTs: A study on the ICT adoption propensity of hotels in Thailand. International Journal of Contemporary Hospitality Management, 17(5), 391–401. S´ anchez-Fernandez, ´ R., Gallarza, M. G., & Arteaga, F. (2020). Adding dynamicity to consumer value dimensions: An exploratory approach to intrinsic values and value outcomes in the hotel industry. International Journal of Contemporary Hospitality Management, 32(2), 853–870. Sarmah, B., Kamboj, S., & Rahman, Z. (2017). Co-Creation in hotel service innovation using smart phone apps: An empirical study. International Journal of Contemporary Hospitality Management, 29(10), 2647–2667. Seo, K., Woo, L., Mun, S. G., & Soh, J. (2021). The asset-light business model and firm performance in complex and dynamic environments: The dynamic capabilities view. Tourism Management, 85, Article 104311. Shin, H., Perdue, R. R., & Kang, J. (2019). Front desk technology innovation in hotels: A managerial perspective. Tourism Management, 74, 310–318. Sigala, M. (2018). New technologies in tourism: From multi-disciplinary to antidisciplinary advances and trajectories. Tourism Management Perspectives, 25, 151–155. Soh, C., & Markus, M. L. (1995). How IT creates business value: A process theory synthesis. ICIS 1995 Proceedings, 4. Tams, S., Grover, V., & Thatcher, J. (2014). Modern information technology in an old workforce: Toward a strategic research agenda. The Journal of Strategic Information Systems, 23(4), 284–304. Williams, T. A., Gruber, D. A., Sutcliffe, K. M., Shepherd, D. A., & Zhao, E. Y. (2017). Organizational response to adversity: Fusing crisis management and resilience research streams. The Academy of Management Annals, 11(2), 733–769. Winter, S. G. (2003). Understanding dynamic capabilities. Strategic Management Journal, 24(10), 991–995. Xie, C., Zhang, J., Chen, Y., & Morrison, A. M. (2022). Hotel employee perceived crisis shocks: Conceptual and scale development. Journal of Hospitality and Tourism Management, 51, 361–374. Yoo, Y., Boland, R. J., Jr., Lyytinen, K., & Majchrzak, A. (2012). Organizing for innovation in the digitized world. Organization Science, 23(5), 1398–1408. Zhang, S., Hu, Z., Li, X., & Ren, A. (2022). The impact of service principal (service robot vs. human staff) on service quality: The mediating role of service principal attribute. Journal of Hospitality and Tourism Management, 52, 170–183. L. Fan et al.

Journal of Hospitality and Tourism Management 57 (2023) 74–83 Available online 14 September 2023 1447-6770/© 2023 The Authors. Published by Elsevier Ltd. on behalf of CAUTHE - COUNCIL FOR AUSTRALASIAN TOURISM AND HOSPITALITY EDUCATION. All rights reserved. How does the usage of robots in hotels affect employees’ turnover intention? A double-edged sword study Lan-Xia Zhang a , Jia-Min Li a,* , Le-Le Wang a , Meng-Yu Mao a , Ruo-Xi Zhang b a School of Business Administration, Northeastern University, Shenyang, Liaoning Province, 110167, China b School of Management, Harbin Institute of Technology, Harbin, Heilongjiang Province, 150001, China ARTICLE INFO Keywords: The usage of robots Turnover intention Conscientiousness Work autonomy Job insecurity Double-edged sword effect ABSTRACT The purpose of this paper is to examine the double-edged sword effect of the usage of robots on hotel employees’ turnover intention. We also examined the mediating role of work autonomy and job insecurity, as well as the moderating role of conscientiousness. Based on conservation of resources theory, we used a time-lagged method to select a sample of 370 employees from 18 hotels in China. We performed statistical analyses using SPSS 23.0 and Mplus 7.4. We found that the usage of robots reduces turnover intention through work autonomy and increases turnover intention through job insecurity. Conscientiousness moderated the effects of the usage of robots on work autonomy and job insecurity, as well as the indirect effects of work autonomy and job insecurity. Our study enriches the research on the usage of robots in the hospitality industry and expands scholars’ exploration of conscientiousness. Our findings change previous scholars’ perceptions of employee’ conscientiousness. It is crucial for hotel managers to acknowledge that the usage of robots is a double-edged sword. Hotel managers need to recruit employees who use robots from a complementary perspective and do not always have to recruit employees with high conscientiousness. 1. Introduction With the rapid development of artificial intelligence, advanced technologies represented by robots have been developing rapidly and quickly gaining popularity in the hotel industry (Choi, Choi, Oh, & Kim, 2020; Li, Wu, Wu, & Goh, 2023). Due to increasing consumer demands, the hotel industry is currently encountering challenges in improving the service quality of its employees. The usage of robots in hotels not only effectively addresses the needs of customers but also brings convenience to the employees. (Fuentes-Moraleda, Diaz-Perez, Orea-Giner, Munoz-Mazon, & Villace-Molinero, 2020). Therefore, many hotels have introduced robots to assist their employees, which has greatly accelerated the popularity of robots in the hospitality industry (Choi et al., 2020; Fuentes-Moraleda et al., 2020; Luo, Vu, Li, & Law, 2021). However, the usage of robots in hotels not only brings convenience to employees, but also brings many problems and challenges. Furthermore, with the growing intelligence of service robots in hotels, a wide range of routine tasks can now be fully carried out independently by robots (Osawa et al., 2017). This progress allows for the reduction in workforce requirements specifically assigned to these tasks, potentially resulting in the displacement of employees in particular positions. As a result, employees without the required skills for alternative roles may feel more job insecurity and struggle to stay focused on their work (Wu, Li, & Wu, 2022). Therefore, the usage of robots in the hospitality industry is a double-edged sword, with both positive and negative effects on employees. Turnover intention is not a new research topic, and scholars have extensively investigated the impact of using robots on this aspect. However, the research findings are inconclusive. While the majority of studies indicate that the usage of robots tends to increase employees’ turnover intention (Khaliq, Waqas, Nisar, Haider, & Asghar, 2022; Koo, Curtis, & Ryan, 2021; Li, Bonn, & Ye, 2019), there are also some studies found that the usage of robots can actually reduce the employees’ turnover intention (Chang et al., 2021). The reason for the inconsistent research conclusions may due to the fact that the usage of robots is a double-edged sword on turnover intention. However, few studies have explored the impact of the usage of robots on turnover intention from the perspective of the double-edged sword. By exploring the double-edged sword effects of the usage of robots on turnover intention, we can better integrate the existing inconsistent * Corresponding author. E-mail addresses: [email protected] (L.-X. Zhang), [email protected] (J.-M. Li), [email protected] (L.-L. Wang), [email protected] (M.-Y. Mao), [email protected] (R.-X. Zhang). Contents lists available at ScienceDirect Journal of Hospitality and Tourism Management journal homepage: www.elsevier.com/locate/jhtm https://doi.org/10.1016/j.jhtm.2023.09.004 Received 30 March 2023; Received in revised form 4 August 2023; Accepted 10 September 2023

Journal of Hospitality and Tourism Management 57 (2023) 74–83 75 research conclusions. On the one hand, it helps us gain a deeper understanding of the complex relationship between these two variables. Traditionally, it is widely believed that introducing robots can improve work efficiency, reduce human errors, and thereby increase employee satisfaction and maintain organizational competitiveness (Fan, Gao, & Han, 2022; Song, Zhang, Hu, & Cao, 2022; Zhang, Hu, Li, & Ren, 2022). However, the usage of robots can also lead to negative effects, such as employees feeling replaced or losing motivation, which may increase turnover intention (Fu, Zheng, & Wong, 2022; Koo et al., 2021; Yu, Shum, Alcorn, Sun, & He, 2022). By studying the double-edged sword effect, we can comprehensively understand the positive and negative impacts of using robots, enabling us to better manage and address this situation. On the other hand, in the hotel and other service industries, the roles and interpersonal interactions of employees play a crucial role in customer satisfaction and service quality. By examining the double-edged sword effect, we can predict the positive and negative reactions of employees to the use of robots, and then take corresponding management measures to improve employee retention and better manage turnover. Therefore, we aim to explore the double-edged sword effects of the usage of robots on turnover intention. The conservation of resources theory (COR) can explain the doubleedged sword effect of the usage of robots on turnover intention. On the one hand, the robots assist employees in completing repetitive and mundane tasks, enabling them to enhance their work autonomy (Yam et al., 2021). On the other hand, the usage of robots in hotels can raise worry among employees. Employees are concerned that their jobs will be replaced by robots when they learn about their capabilities, thus increasing job insecurity (Koo et al., 2021; Wu et al., 2022). Therefore, we selected work autonomy and job insecurity as mediating variables to examine the mediating role of work autonomy and job insecurity between the usage of robots and turnover intention. Complementary theory suggests that people prefer to maintain a harmonious and balanced complementary match between their own attributes and those of other work entities (Heider, 1982). Conscientiousness is a trait of employees who are conscientious and careful (Bogg & Roberts, 2004; Li, Zhang, Zhang, & Zhang, 2023). Robots, as artificial intelligences, are fully controlled by the operating instructions of employees, and they also demonstrate conscientiousness and meticulousness in their work. The characteristics of a conscientious employee and robots share similarities. Therefore, we want to explore the moderating effect of conscientiousness. 2. Theoretical foundation and hypothesis development 2.1. Theoretical foundation COR argues that individuals actively strive to acquire, value and protect their personal resources. Individuals feel stressed when their critical resources are threatened, lost or even when they cannot continue to obtain new resources (Hobfoll, 1989). COR focuses on the idea that individuals’ behavior is based on acquiring and preserving the resources they need. These resources not only alleviate stress but also fulfill individuals’ immediate survival and future needs (Hobfoll, 1989; Hobfoll, Halbesleben, Neveu, & Westman, 2018). A large number of studies based on the double-edged sword perspective currently use COR as a theoretical basis (Khaliq et al., 2022). In this study, the usage of robots in the hotel can be considered as a resource gain for certain employees, as the robots assist the hotel employees in many tasks (Fan et al., 2022; Song et al., 2022). However, for another part of the employees, the usage of robots represents a loss of resources, posing a threat to their roles and responsibilities (Fu et al., 2022; Koo et al., 2021). Therefore, it is appropriate for this study to explore the double-edged sword effect of the usage of robots in hotels, employing COR as a theoretical basis. 2.2. The usage of robots and turnover intention With the development of artificial intelligence, robots are widely used in the hospitality industry, which has implications for organizations, employees and customers. Scholarly research on the usage of robots and artificial intelligence is on the rise (Budhwar, Malik, De Silva, & Thevisuthan, 2022; Vrontis et al., 2022). At the macro level, the usage of robots will reduce the proportion of employed people, replace some lowand middle-skilled jobs, and widen the social gap (Lu et al., 2020; Wu, Li, Wang, & Zhang, 2023). At the meso level, the usage of robots will increase labor productivity and improve business economic efficiency (Cheng, Jia, Li, & Li, 2019). At the micro level, the usage of robots affects the business management model and increases the job insecurity of employees (Koo et al., 2021; Wu et al., 2022). Thus, the usage of robots is a double-edged sword that can have both positive and negative effects on the organization and its employees. Turnover intention refers to an employee’s willingness and tendency to voluntarily terminate his or her membership in the organization (Bothma & Roodt, 2013). Turnover intention and turnover behavior are two stages that describe the turnover of employees, and employees go through a complex process from turnover intention to turnover behavior (Cho & Lewis, 2012; Manolopoulos, Peitzika, Mamakou, & Myloni, 2022). Previous studies have explored the antecedent variables and mechanisms of turnover intention in the hospitality industry from many perspectives (Popa, Lee, Yu, & Madera, 2023; Pu, Ji, & Sang, 2022; Yin, Bi, & Ni, 2022). While the usage of robots can influence the psychology of employees when they leave a job, the transition from this willingness to actual behavior can be influenced by other factors. These factors are not determined by the usage of robots (Cheng et al., 2019; Choi et al., 2020). Therefore, we focus on the impact of the usage of robots on the turnover intention of hotel employees. In general, the usage of robots will reduce the turnover intention of employees (Li et al., 2019). This is because robots can assist employees perform many tedious, tiring and repetitive tasks at work. With the help of robots, hotel employees can spend more time doing those mental tasks, freeing up their workforce (Lu et al., 2020; Osawa et al., 2017). Based on the COR, the usage of robots can be considered as a pool of resources that replenishes employees with additional resources. When employees have access to these resources, they are more likely to exhibit lower turnover intention (Hobfoll, 1989; Hobfoll et al., 2018). Therefore, the usage of robots in this case effectively reduces the employees’ turnover intention. However, previous studies have found that many hotel employees do not quickly adapt to working with robots. In addition to fulfilling their job responsibilities, this group of hotel employees must invest time in learning the operation and usage requirements of the robots (Cheng et al., 2019; Choi et al., 2020; Fuentes-Moraleda et al., 2020). Because of the diverse educational backgrounds and age range of frontline hotel staff, acquiring the skills to operate robots is a challenging task for many employees. This not only adds to their workload but also raises concerns about job security (Nam, 2019). Based on the COR, the usage of robots depletes employees’ resources, which, over time, can increase their turnover intention (Hobfoll, 1989; Hobfoll et al., 2018). Therefore, the usage of robots in this case is associated with an increase in employees’ turnover intention. 2.3. The mediating role of work autonomy The usage of robots in the hospitality industry has had a huge impact on the industry. For example, many hotels are now employing robots to assist staff in various tasks, including customer service and answering inquiries in the hotel lobby, leading to enhanced customer experiences (Cheng et al., 2019; Choi et al., 2020; Lu et al., 2020). Work autonomy refers to the degree of independence employees have in determining their tasks, prioritizing their work, and choosing their work methods (Gardell, 1977). Currently, there are various types of robots in hotels, L.-X. Zhang et al.

Journal of Hospitality and Tourism Management 57 (2023) 74–83 76 such as check-in robots, room delivery robots, concierge robots and food delivery robots, which can help employees perform many tasks. When employees use robots to assist them in their work, they can decide the content and sequence of the robot’s work according to their actual situation. Based on the COR, the usage of robots is considered as a valuable resource pool that provides additional resources for hotel employees. This increase in resources leads to a higher sense of work autonomy among employees (Hobfoll, 1989). Therefore, the usage of robots positively affects work autonomy. Hotel employees with high work autonomy can effectively incorporate the assistance of robots into their work schedules, thus reducing turnover intention (Jaiswal & Dhar, 2017). Work autonomy provides employees with opportunities for development and growth. When employees’ work autonomy is restricted, they may worry about limitations on their career growth and promotion opportunities, which may prompt them to seek other employment opportunities that offer more avenues for development, thereby increasing their turnover intention. According to COR, individuals are driven to protect and maintain their personal resources, including work autonomy (Hobfoll, 1989). Hotel employees, in particular, are inclined to stay in their current positions in order to protect their resources, especially their work autonomy. If employees perceive a threat or reduction in their work autonomy, they may experience pressure related to resource competition and increase their turnover intention (Galletta, Portoghese, & Battistelli, 2011). Therefore, work autonomy negatively affects turnover intention. In conclusion, based on the fundamental framework of the COR, this study considers work autonomy as an important mechanism in the relationship between the usage of robots and employees’ turnover intention. In hotels where robots are utilized, the assistance provided by robots can serve as an additional resource for hotel employees, thereby increasing their level of work autonomy. The enhancement of work autonomy, in turn, allows employees to access further additional resources, thereby reducing their turnover intention. Therefore, this study posits that work autonomy serves as a mediating variable between the usage of robots and employees’ turnover intention in the hotel industry. Thus, we propose: Hypothesis 1. Work autonomy mediates the relationship between the usage of robots and turnover intention. 2.4. The mediating role of job insecurity Job insecurity is an employee’s perception of a threat to their job, and it reflects the state in which their resources are being depleted (Greenhalgh & Rosenblatt, 1984). Previous studies have found that job insecurity can cause employees to have negative emotions such as anxiety and complaints (Jordan, Ashkanasy, & Hartel, 2002; Witte, 1999). After employees start using robots, they may perceive a threat to their personal resources, which can lead to feelings of insecurity. They may worry that the usage of robots could result in their positions being replaced or a decrease in job demand, leading to a competition for resources. This resource competition increases employees’ sense of insecurity (Koo et al., 2021; Wu et al., 2022). In addition, the introduction of robots can potentially change employees’ job roles and responsibilities. Employees may feel a decrease in control over their work and uncertainty about their position and value within the team. This can lead to a depletion of their resources and subsequently increase their feelings of job insecurity. Based on the COR, individuals strive to protect and maintain their personal resources (Hobfoll, 1989). In the hotel industry, employees using robots may experience job insecurity due to the depletion of their resources. Therefore, the ueage of robots positively affects job insecurity. When employees experience job insecurity, they may perceive a threat and depletion of their personal resources. They are concerned that the usage of robots could result in their positions being replaced or a decrease in job demand, leading to a competition for resources (Lee & Jeong, 2017). Based on the COR, this competition for resources further intensifies employees’ job insecurity, motivating them to seek alternative job opportunities or consider leaving their current positions. Therefore, job insecurity positively affects turnover intention. In summary, based on the fundamental framework of the COR, this research suggests that job insecurity is another important mechanism through which the usage of robots influences turnover intention. In hotels where robots are introduced, hotel employees may experience pressure related to operating intelligent devices and the potential threat of being replaced. These factors can lead to a loss of resources for hotel employees, thereby increasing their job insecurity. As job insecurity intensifies, employees further deplete their personal resources, subsequently elevating their turnover intention. Therefore, this study proposes that job insecurity serves as a mediating variable between the usage of robots and turnover intention in the hotel industry. Thus, we propose: Hypothesis 2. Job insecurity mediates the relationship between the usage of robots and turnover intention. 2.5. The moderating role of conscientiousness In the last century, many disciplines emphasized the importance of conscientiousness. Especially in the workplace, conscientiousness was considered a key factor in predicting performance (Bogg & Roberts, 2004; Li, Zhang, et al., 2023). Traditional intelligent devices only required simple operations to function, where employees input commands and the devices executed them. At that time, employees with a sense of responsibility operating these traditional devices were regarded as advantageous, and these devices were favored by conscientious employees (Menter, 1973). Since the beginning of the 21st century, technology and intelligent machines have undergone significant changes. Compared to the intelligent machines of the previous century, current intelligent machines not only alleviate the burden of repetitive tasks for employees but also have the capability to make decisions through learning (Lee, Kusbit, Metsky, & Dabbish, 2015). There are distinct differences between current intelligent machines and traditional ones. Therefore, when recruiting employees to collaborate with artificial intelligence, a company’s perspective on employee conscientiousness may undergo changes. While highly conscientious employees can efficiently utilize technology to accomplish current tasks, these viewpoints are based on conclusions drawn from technology of the previous century. This study posits that 21st-century intelligent machines might challenge this consensus due to the lack of complementarity between these machines and conscientious employees. Tang et al. (2021) has already demonstrated this perspective, and to further validate this finding, we employ the complementary theory to explain this assertion. Complementary theory suggests that people prefer complementary matches that maintain a balance between their own attributes and the attributes of other working entities. It is often applied to interactions between people, focusing on the match between one person’s personality traits and another person’s submissive traits (Heider, 1982; Kiesler, 1983). With the spread of robots in the hospitality industry, employees are working with a shift from people to robots (Cheng et al., 2019). Employees with conscientiousness will show self-discipline and carefulness at work (Schmidt-Wilk, Alexander, & Swanson, 1996). This is consistent with the characteristics of robots at work. According to complementarity theory, employees prefer to work with employees whose attributes are complementary to their own. However, this preference can hinder the efficiency of employees with high conscientiousness when working with robots (Heider, 1982; Kiesler, 1983). Based on this, this study suggests that conscientiousness can moderate the effects of the usage of robots on work autonomy and job insecurity. Based on the complementary theory, employees with high conscientiousness will demonstrate a strong sense of self-discipline and conscientiousness at work. They dedicate a significant amount of time L.-X. Zhang et al.

Journal of Hospitality and Tourism Management 57 (2023) 74–83 77 and energy to their tasks. Interestingly, intelligent machines in the workplace exhibit similar characteristics to those of highly conscientious employees, such as standardization and intelligence (Bogg & Roberts, 2004; Tang et al., 2021). Therefore, the autonomy and systematic nature of intelligent machines resemble the characteristics of conscientious employees, creating a mismatch. Specifically, conscientious employees may be better suited to work with traditional intelligent devices, as they may perceive the information and suggestions provided by intelligent machines as untrustworthy. In other words, they are not guided or constrained by intelligent machines and may prefer to develop their own insights and approaches to accomplish tasks. As a result, these employees may derive fewer idle resources from interacting with intelligent machines. Unfortunately, this further restricts the positive impact of using robots on job autonomy, thereby diminishing the advantages of utilizing intelligent technology (Heider, 1982; Kiesler, 1983). At the same time, employees with high conscientiousness are more likely to feel threatened by robots in relation to their own jobs. They perceive that robots possess similar characteristics to themselves and have the capability to completely replace their roles, thereby intensifying their sense of job insecurity (Xu, Zhu, & Li, 2022). Conversely, employees with low conscientiousness are not very disciplined and responsible at work, often getting easily distracted by external factors (Schmidt-Wilk et al., 1996). These attributes align with the characteristics of robots, thus creating a complementary match. Based on the complementarity theory, Employees with lower conscientiousness are more accepting of intelligent machines and consider them as good partners in their work. These employees have lower requirements for controlling the work process and are more likely to perceive intelligent machines as trusted allies. In their collaboration with intelligent machines, employees with lower conscientiousness are in a more advantageous position, allowing them to fully leverage the advantages of intelligent machines. Compared to employees with higher conscientiousness, they perceive intelligent machines as more valuable. As a result, these employees make better use of and gain greater access to the idle resources derived from interacting with intelligent machines, thereby strengthening the positive impact of using robots on work autonomy (Bogg & Roberts, 2004; Tang et al., 2021). At the same time, they do not take their jobs very seriously and are not overly concerned that the usage of robots will replace their jobs. Therefore, this group of employees does not feel a high level of job insecurity (Heider, 1982; Kiesler, 1983). Thus, we propose: Hypothesis 3a. Conscientiousness moderates the positive relationship between the usage of robots and work autonomy, such that this positive relationship is stronger at lower levels of conscientiousness. Hypothesis 3b. Conscientiousness moderates the positive relationship between the usage of robots and job insecurity, such that this positive relationship is weaker at lower levels of conscientiousness. Further, conscientiousness not only moderates the relationship between the usage of robots and work autonomy, but also moderates the relationship between the usage of robots and job insecurity. Work autonomy and job insecurity mediated the effect of the usage of robots on turnover intention. We suggest that conscientiousness also moderated the indirect effects of work autonomy and job insecurity. Based on the Hypothesis 1 and Hypothesis 3a, the characteristics of robots at work are similar to those of employees with high conscientiousness, so employees who work with robots will not perceive a significant improvement in work autonomy when employees with high conscientiousness. Work autonomy is expected to reduce turnover intention. Therefore, when employees with high conscientiousness, employees who work with robots are less likely to perceive an increase in work autonomy, and therefore increase their turnover intention. In other words, the higher the employee’s conscientiousness, the weaker the indirect effect of work autonomy in the usage of robots and turnover intention. Based on the Hypothesis 2 and Hypothesis 3b, the characteristics of robots at work are similar to those of employees with high conscientiousness, so employees who work with robots will perceive a significant improvement in job insecurity when employees with high conscientiousness. Job insecurity is expected to increase turnover intention. Therefore, when employees with high conscientiousness, employees who work with robots are more likely to perceive an increase in job insecurity, and therefore increase their turnover intention. In other words, the higher the employee’s conscientiousness, the stronger the indirect effect of job insecurity in the usage of robots and turnover intention. Thus, we propose: Hypothesis 4a. The indirect effect of usage of robots on turnover intention via work autonomy is moderated by conscientiousness, such that the indirect effect will be weaker when conscientiousness is higher. Hypothesis 4b. The indirect effect of usage of robots on turnover intention via job insecurity is moderated by conscientiousness, such that the indirect effect will be stronger when conscientiousness is higher. The model diagram of this study is shown in Fig. 1. 3. Methods 3.1. Pilot test Since our questionnaire was translated from an English scale, we conducted a pilot test to assess the quality of the Chinese version of the questionnaire. In August 2022, we established contact with a hotel in Shenyang, Liaoning Province, China, that met the requirements for our survey through alumni connections. With the assistance of the hotel manager, we conducted the pilot test for our study. Ultimately, we collected 147 valid questionnaires. To validate the quality of the questionnaire, we conducted reliability analysis and validity analysis. The results of the independent t-test showed that all of the designed questions reached the levels of significance that were necessary for discrimination. Furthermore, the Cronbach’s α for the usage of robots, job insecurity, work autonomy, turnover intention, and conscientiousness were 0.921, 0.892, 0.832, 0.914, and 0.867, respectively. This indicates that the data has good reliability. To access the validity of the study, the Kaiser-Meyers-Olkin (KMO) was used. The validity of this questionnaire was found to be in the acceptable range, which indicate high construct validity (Kaiser, 1974). The results indicated that the questionnaire data passed all of these tests. This suggests that the translated Chinese version of the questionnaire exhibits good quality and is suitable for formal research. 3.2. Participants and procedure In Liaoning province, China, service employees, represented by the hospitality industry, are already widely equipped with various types of service robots at work. Therefore, the participants in this study were Fig. 1. Research model. L.-X. Zhang et al.

Journal of Hospitality and Tourism Management 57 (2023) 74–83 78 frontline employees from 18 hotels in Shenyang, Liaoning Province. We followed the following criteria when selecting hotels, i.e., all hotels must already be using robots to assist their employees. Although the target hotel was equipped with robots for employees’ work, in order to ensure the accuracy of the questionnaire results and to filter out irrelevant participants, screening questions were included in the questionnaire to confirm whether the robots were used in the workplace. Participants who responded “No” were deemed ineligible, and their questionnaires were excluded from the analysis to maintain data accuracy. Researchers contacted hotel managers for their assistance in the study, and questionnaires were subsequently distributed to frontline employees across different hotel functions through the hotel managers’ facilitation. In order to reduce the effect of common method bias on the relationship between variables, the following measures were taken: (1) Data were collected in two stages, each at an interval of one month. (2) Anonymous responses were used to reduce participants’ tendency to personal bias. (3) The names of the variables were not displayed in the questionnaire to hide the research items and to ensure that participants answered the questions according to their true personal feelings. The questionnaire was distributed at two different time points in this study. To ensure data consistency, employees were requested to provide the last four digits of their cell phone number in both questionnaires. The first questionnaire was distributed on September 10, 2022, aimed to measure the control variables, the usage of robots and job insecurity. 460 questionnaires were distributed and 420 questionnaires were returned, with a return rate of 91.304%. The second phase study began on October 10, 2022, with the distribution of questionnaires to measure conscientiousness, work autonomy, and turnover intention. A total of 420 questionnaires were distributed and 382 questionnaires were returned, with a return rate of 90.952%. The questionnaires were sorted and screened, and 370 valid questionnaires were obtained after deleting those questionnaires with obvious patterns of responses and those with missing answers. In the valid sample, male employees accounted for 53.514% and female employees accounted for 46.486%. The largest number of hotel employees have high school or junior college education, accounting for 33.514%. The respondents reported an average age of 36.522 years, and an average organizational tenure of 15.365 years. In addition, the majority of the sample belonged to service and sales employees who used service robots to assist themselves in cleaning and delivering items needed by customers, which accounted for 68.648%. Another part of the sample belonged to technicians and associate professionals who needed robotic assistance in transporting tools at work, which accounted for 22.162%. Finally, there are other frontline employees who use robots, accounting for 9.189%. 3.3. Measures We translated an English questionnaire into Chinese using a translation-back translation procedure (Brislin, 1970). We invited two graduate students from the Foreign Language Institute for this task. One student translated the questionnaire from English to Chinese, while the other student translated it back from Chinese to English. Finally, the author, two professors specializing in tourism management and human resources management, and two doctoral students engaged in a discussion to address any inconsistencies. Through this collaborative effort, a consensus was reached, and the survey questionnaire was formed. 3.3.1. The usage of robots We measured this variable using a 3-item scale developed by Medcof (1996), drawing on Tang et al. (2021). This scale is a 5-point Likert-type scale, with 1 representing “strongly disagree” and 5 representing “strongly agree”. A sample question item is “I used robot to carry out most of my job functions.” The Cronbach’s α for this scale was 0.855. 3.3.2. Work autonomy Work autonomy were assessed with Spreitzer’s (1995) three-item scale. This scale used a 5-point Likert-type scale, strongly disagree (1) to strongly agree (5) sacle format. Sample items include “I have significant autonomy in determining how I do my job.” The Cronbach’s α for this scale was 0.800. 3.3.3. Job insecurity Mauno and Ulla (2001) developed scale was used to measure job insecurity. This scale is a 5-point Likert-type scale, with 1 representing “strongly disagree” and 5 representing “strongly agree”. Sample items include “Your job is insecure.” The Cronbach’s α for this scale was 0.858. 3.3.4. Turnover intention Mobley, Griffeth, Hand, and Meglino (1979) developed sacle was used to measure turnover intention. This scale is a 5-point Likert-type scale, with 1 representing “strongly disagree” and 5 representing “strongly agree”. Sample items include “I basically have no desire to leave this current business.” The Cronbach’s α for this scale was 0.851. 3.3.5. Conscientiousness Conscientiousness was measured with Pathki, Kluemper, Meuser, and Mclarty’s (2022) four-item scale. This scale is a 5-point Likert-type scale, with 1 representing “strongly disagree” and 5 representing “strongly agree”. Sample items include “At work, I get my tasks done right away.” The Cronbach’s α for this scale was 0.817. 3.3.6. Control variable Scholars (Alkahtani, 2015; Belete, 2018) found that age, gender, education and tenure affect turnover intention. In addition, since our participants were from 18 different hotels, we controlled for these hotels to avoid organizational differences. Specifically, we used organization size and organization type as control variables. Therefore, the above variables were used as control variables in this study. 4. Results We performed statistical analyses using SPSS 23.0 and Mplus 7.4. We used SPSS 23.0 for descriptive statistical analysis, correlation analysis, reliability analysis and validity analysis. And we used Mplus 7.4 for confirmatory factor analysis and hypothesis testing. 4.1. Common method bias To test the data of this study for serious common method bias, we used Harman’s One-Factor Test for validation (Podsakoff & Organ, 2016). We performed exploratory factor analysis on all measured items simultaneously, and the unrotated factor analysis showed that the first principal component explained 30.799% of the total variance. The results of the data analysis indicate that there is no serious common method bias in the data of this study. In addition, given that the results of the Harman’s One-Factor Test may not be sensitive, this study added the unmeasured latent method factor to the five-factor model. Subsequently, the model was compared with the five-factor model, and it was found that there was little change in the indicators (ΔCFI = 0.017, ΔTLI = 0.015, and ΔRMSEA = 0.011), which again indicates that the problem of common method bias in this study is not serious (Podsakoff, MacKenzie, & Podsakoff, 2012). 4.2. Descriptive statistical analysis We used SPSS 23.0 for correlation analysis, reliability analysis and validity analysis and the results are shown in Table 1. The composite reliability of each variable with the square root of AVE is also shown in Table 1. It can be seen that the composite reliability of each variable is greater than 0.850, which indicates that the study data has good L.-X. Zhang et al.

Journal of Hospitality and Tourism Management 57 (2023) 74–83 79 composite reliability. The square root of AVE of each variable is greater than the correlation coefficient of the variable with other variables, which indicates that the study data have good discriminant validity. In addition, Table 2 shows the factor loadings of the variables. As can be seen from Table 2, the factor loadings of all the question items have a value greater than 0.75. This also indicates that the quality of the data is good. 4.3. Confirmatory factor analysis We conducted confirmatory factor analysis using Mplus 7.4 on the usage of robots, work autonomy, job insecurity, conscientiousness, and turnover intention and further tested the discriminant validity of each variable, and the results were shown in Table 3. It can be seen that the five-factor model has the best data fit compared with other factorial models (χ2 /df = 1.323, RMSEA = 0.030,GFI = 0.951,CFI = 0.984,NFI = 0.939,NNFI = 0.981,TLI = 0.981,AGFI = 0.934,IFI = 0.985). The results indicate that the data passed the confirmatory factor analysis. 4.4. Hypotheses testing We used Mplus 7.4 to construct structural equation model to test the hypothesis, the model has a good fit index (χ2 /df = 2.433, RMSEA = 0.029,GFI = 0.921,CFI = 0.953,NFI = 0.92). The coefficients and significance of each path were shown in Fig. 2. It can be seen that the usage of robots was negatively related to turnover intention (β = − 0.261, SE = 0.043, p < 0.001), the usage of robots was positively related to job insecurity (β = 0.283, SE = 0.038, p < 0.001) and work autonomy (β = 0.233, SE = 0.044, p < 0.001), work autonomy (β = − 0.186, SE = 0.050, p < 0.001) and job insecurity (β = 0.274, SE = 0.055, p < 0.001) were negatively related to turnover intention. To test the mediating role of work autonomy and job insecurity, we tested for mediation using the conditional indirect effects approach (Bootstrap = 5000). We first examined the mediating effect of work autonomy and found the indirect effect estimate to be − 0.043, with a 95% confidence interval not containing 0 (95% Boot CI = [− 0.086, − 0.017]). Thus, work autonomy mediated between the usage of robots and turnover intention. Then, we examined the mediating effect of job insecurity. We found the indirect effect estimate of 0.077, with a 95% confidence interval not containing 0 (95% Boot CI = [0.039, 0.084]). Thus, job insecurity mediated between the usage of robots and turnover intention. H1 and H2 were supported. We created the latent interaction of sequential breaches by using the “XWITH” option in Mplus. Research indicates this approach can produce more accurate estimates of the interaction effect (Bamberger & Table 1 Mean. standard deviation and correlation coefficient of each variable. Mean SD 1 2 3 4 5 6 7 8 9 10 11 1.Gender 1.465 0.499 – 2.Age 36.522 8.136 0.044 – 3.Education 2.170 1.072 − 0.032 0.049 – 4.Tenure 15.365 8.199 0.050 0.981*** − 0.133* – 5. Oganization size 2.124 0.672 − 0.067 0.034 0.026 0.035 – 6. Organization type 2.445 0.782 0.036 − 0.108 0.053 0.009 0.004 – 7.The usage of robots 3.631 1.035 0.027 − 0.040 0.045 − 0.049 0.234* 0.028 0.855 8.Job insecurity 3.978 0.808 0.081 − 0.056 0.063 − 0.067 0.120* 0.019* 0.368*** 0.858 9.Turnover intention 2.169 0.922 − 0.008 0.122* 0.229*** 0.073 − 0.021 0.028 − 0.283*** 0.334*** 0.851 10.Work autonomy 3.955 0.900 − 0.022 − 0.094 0.056 − 0.102 0.027** 0.031 0.272*** 0.272*** − 0.285*** 0.800 11.Conscientiousness 3.743 0.862 0.066 0.054 0.160** 0.030 0.042 0.021** 0.286*** 0.335*** − 0.260*** 0.312*** 0.817 Square root of AVE – – – – – – – – 0.881 0.799 0.832 0.840 0.804 CR – – – – – – – – 0.912 0.899 0.900 0.878 0.880 The values on the diagonal are Cronbach’s α of each variable. CR = composite reliability AVE = average variance extracted. *p < 0.05 **p < 0.01 ***p < 0.001. Table 2 Factor loadings for all scale items. Variables Items Factor Loading The usage of robots A1 0.881 A2 0.882 A3 0.879 Work autonomy B1 0.839 B2 0.842 B3 0.840 Job insecurity C1 0.774 C2 0.782 C3 0.828 C4 0.818 C5 0.794 Turnover intention D1 0.847 D2 0.830 D3 0.818 D4 0.832 Conscientiousness E1 0.817 E2 0.802 E3 0.811 E4 0.785 L.-X. Zhang et al.

Journal of Hospitality and Tourism Management 57 (2023) 74–83 80 Belogolovsky, 2017; Cheung & Lau, 2017). As seen in Fig. 2, conscientiousness moderated the relationship between the usage of robots and work autonomy (β = − 0.142, SE = 0.044, p < 0.01). H3a was supported. Conscientiousness also moderated the relationship between the usage of robots and job insecurity (β = − 0.134, SE = 0.038, p < 0.001). H3b was partially supported. To further test H3, we used the simple slope analysis to plot the moderating effect of conscientiousness between the usage of robots and work autonomy, and job insecurity. The moderating role diagrams were shown in Figs. 3 and 4. Using simple slope analysis, we found that when conscientiousness moderates the relationship between the usage of robots and work autonomy, highly conscientious employees have higher work autonomy overall than lowly conscientious employees. As expected, the slope of the line between the usage of robots and work autonomy was greater for those with lower levels of conscientiousness, compared to those with higher levels of conscientiousness (Fig. 3). H3a was further supported. Meanwhile, when conscientiousness moderated the relationship between the usage of robots and job insecurity, highly conscientious employees have higher job insecurity overall than lowly conscientious employees. Contrary to our expectations, the slope of the line between the usage of robots and job insecurity was greater for those with lower levels of conscientiousness, compared to those with higher levels of conscientiousness (Fig. 4). H3b was partially supported. Finally, we examined the moderated mediation model. Table 4 illustrated results of conditional indirect effects. The conditional indirect effect between the usage of robots and turnover intention via work autonomy was negative and significant at lower levels of conscientiousness (estimate = − 0.049, 95% CI = [− 0.090, − 0.017]), but not at higher levels (estimate = -0.004, 95% CI = [− 0.025, 0.017]); the difference between these effects was significant. Meanwhile, the indirect effect of work autonomy was stronger when conscientiousness was lower. Therefore, H4a was supported. Similarly, the conditional indirect effect between the usage of robots and turnover intention via job insecurity was negative and significant at lower levels of conscientiousness (estimate = − 0.088, 95% CI = [− 0.145, − 0.039]), but not at higher levels (estimate = − 0.024, 95% CI = [− 0.054,0.002]); the difference between these effects was significant. However, the indirect effect of job insecurity was stronger when conscientiousness was lower. Therefore, H4b was partially supported. 5. Discussion The usage of robots in hotels has gained a lot of attention from scholars. Scholars have explored this topic and have examined the impact of robotics on employees and organizations from both positive and negative perspectives (Fu et al., 2022; Lee, Lee, & Kim, 2021; Wang, Ho, Yeh, & Huan, 2022). This study constructs a double-edged model of the effect of the usage of robots on turnover intention in hotels and explores the mediating role of work autonomy and job insecurity, as well as the moderating role of conscientiousness. We found that conscientiousness moderated the effect of robot use on job insecurity and work Table 3 The results of confirmatory factor analysis. Model χ2 /df RMSEA GFI CFI NFI NNFI TLI AGFI IFI 1. Five-factor model (including 1, 2, 3, 4,5) 1.323 0.030 0.951 0.984 0.939 0.981 0.981 0.934 0.985 2. Four-factor model (including 1 + 2, 3, 4,5) 4.018 0.090 0.842 0.850 0.811 0.824 0.824 0.794 0.851 3. Three-factor model (including 1 + 2+3, 4,5) 7.066 0.128 0.720 0.692 0.661 0.646 0.646 0.642 0.694 4. Two-factor model (including 1 + 2+3 + 4,5) 8.696 0.144 0.674 0.604 0.577 0.551 0.551 0.589 0.607 5. One-factor model (including 1 + 2+3 + 4+5) 10.765 0.163 0.621 0.494 0.473 0.431 0.431 0.526 0.497 1 = The usage of robots, 2 = Job insecurity, 3 = Turnover intention, 4 = Work autonomy, 5 = Conscientiousness. Fig. 2. Path coefficients of the model (Note: ** indicates p < 0.01, ***indicates p < 0.001). Fig. 3. Moderating effect of conscientiousness on the relationship between the usage of robots and work autonomy. L.-X. Zhang et al.

Journal of Hospitality and Tourism Management 57 (2023) 74–83 81 autonomy. For employees with either high or low conscientiousness, the increase in the usage of robots would increase work autonomy. As expected, the slope of the line between the usage of robots and work autonomy was greater for those with lower levels of conscientiousness, compared to those with higher levels of conscientiousness. This result is consistent with the finding of Tang et al. (2021) and with our hypothesis. However, for employees with either high or low conscientiousness, the increase in the usage of robots would increase job insecurity. Contrary to our expectations, the slope of the line between the usage of robots and job insecurity was greater for those with lower levels of conscientiousness, compared to those with higher levels of conscientiousness. The reason for this result may be that the usage of robots in hotels has a different focus on work autonomy and job insecurity. Specifically, employees with conscientiousness are aware that their work is not easily replaced by robots. Since these employees are conscientious, even if robots may replace some employees, they will not replace employees with conscientiousness. 5.1. Theoretical implications Firstly, our research validates COR and extends the complementary theory to a certain extent. With the popularity of artificial intelligence, many scholars have explored the impact of the usage of AI on employee behavior from a resource perspective and have achieved fruitful results (Fan et al., 2022; Song et al., 2022). Building upon previous research, we once again validate the impact of the usage of robots in hotels on turnover intention from a resource perspective. Therefore, our study validates COR. Additionally. Previous studies using complementarity theory have focused on human-human interactions (Grijalva & Harms, 2014; Hu & Judge, 2017; Wu, Zhang, & Li, 2023). However, with the rapid development of technology, many hotels are using robots to assist employees. Cooperation and interaction between humans and robots have become an important issue in theory and practice (Li, Wu, Wu, & Goh, 2023). This study applies the complementarity theory to the research on human-machine interaction, expanding the research perspective of complementarity theory from human-human interaction to human-machine interaction. Therefore, our study also extends the complementarity theory to a certain extent. Secondly, we constructed positive and negative paths to examine the impact of the usage of robots on turnover intention. Scholars have now examined the positive and negative effects of AI on hotel employees separately, and they have come up with conflicting or inconsistent results (Vatan & Dogan, 2021; Yu et al., 2022). The existing studies neglect other theoretical perspectives and rely on specific hidden assumptions. Our study remedies the one-sidedness of considering only the positive or negative effects of robot use in hotels by revealing a more complete relationship and mechanism of action between the usage of robots and turnover intention. We remedy the problem of one-sidedness in single-sided studies and respond to the inconsistency of findings in previous studies. Finally, we have enriched and expanded the research on conscientiousness. Most of the previous studies concluded that employees with conscientiousness perform better at work (Bogg & Roberts, 2004; Li, Zhang, et al., 2023). Some scholars based on other theoretical perspectives have found that employees with conscientiousness are not suitable to work with intelligent machines (Tang et al., 2021). Based on complementary theory, we regard conscientiousness as an important boundary condition. We found that for employees with either high or low conscientiousness, the increase in the usage of robots would increase job insecurity perceptions. But for the high conscientiousness group, the positive impact was bigger. The same moderation effect of conscientiousness on work autonomy. Our findings test and enrich previous scholarly research on conscientiousness. 5.2. Practical implications Firstly, hotel employees and managers need to be aware of both the positive and negative effects of robots. For hotel managers, managers who do not have a plan to fully promote robots in hotels do not effectively motivate employees and sometimes even do the opposite of what they want. Hotel managers need to promote the usage of robots in key departments according to the actual situation of the hotel. At the same time, they should strengthen the training for employees to quickly adapt to the work mode of human-robot collaboration and assign those who are not comfortable working with robots to traditional positions. This will reduce employees’ turnover intention. Fig. 4. Moderating effect of conscientiousness on the relationship between the usage of robots and job insecurity. Table 4 The results of conditional indirect effects. Path Estimate SE 95% CI The usage of robots → Job insecurity → Turnover intention (Mean-1SD) − 0.088 0.027 [-0.145, − 0.039] The usage of robots → Job insecurity → Turnover intention (Mean+1SD) − 0.024 0.014 [-0.054, 0.002] The usage of robots → Work autonomy → Turnover intention (Mean-1SD) − 0.049 0.019 [-0.090, − 0.017] The usage of robots →Work autonomy → Turnover intention (Mean+1SD) − 0.004 0.010 [-0.025, 0.017] L.-X. Zhang et al.

Journal of Hospitality and Tourism Management 57 (2023) 74–83 82 Secondly, managers need to be aware of the changes that the usage of robots brings to employees’ work autonomy and job insecurity. The usage of robots in hotels can reduce the turnover intention through work autonomy and increase the turnover intention through job insecurity. Therefore, on the one hand, hotel managers should try to increase employees’ work autonomy. For example, hotel managers can clarify in their daily work which tasks can be mainly done by robots and let employees know which tasks they can fully decide, so that employees can continuously improve their work autonomy in the process of using robots. On the other hand, hotel managers should even alleviate employees’ job insecurity. For example, managers can organize frequent training, provide necessary work support, etc. To relieve employees’ work pressure, and give employees more work guidance to shape the organization’s fault-tolerant atmosphere to reduce employees’ job insecurity. Finally, managers need to properly understand conscientiousness. In the management practices of hotels, many hotel managers overlook the investigation of employee traits during recruitment. In fact, employee traits have certain reference value in job allocation. Therefore, hotel managers should pay attention to the investigation of employee traits in future recruitment processes. In addition to this, managers should still promote employees to develop conscientiousness in their future management. In the assignment of positions, employees with high conscientiousness mainly perform tasks that are detailed or where robots assist themselves with less work. The employees with average conscientiousness are mainly engaged in those tasks that the robot assists them to complete with a high workload. It is important to note that while this study found complementarity between conscientiousness and robot operation, it does not advocate that companies ignore employee conscientiousness, but rather assign those employees who are dedicated to their jobs to more appropriate positions. 5.3. Limitations and future research directions Firstly, the research sample in this study was employees who use robots in hotels. We did not further classify the types of robots in the study. In management practice, robots in hotels can be classified into various types depending on the nature of their work, and the effect of different types of robots on employees’ turnover intention may also vary. Therefore, the impact of different types of robots on hotel employees’ turnover intention can be further explored in the future. Secondly, although this study used a time-lagged approach to collect data and took many measures to reduce the impact of potential common method bias on the relationship between variables. However, all questionnaires were filled out by hotel employees, which is a single source and also causes potential common method bias. In future studies, hotel managers could be invited to participate in the survey and fill out the questionnaires to effectively reduce the impact of common method bias. Thirdly, this study selected work autonomy, job insecurity, and conscientiousness as important variables in the mechanism of the effect of the usage of robots on turnover intention. However, it is important to acknowledge that there may be other variables influencing this relationship. Therefore, future studies can consider exploring additional key variables to uncover alternative mechanisms through which robot usage affects employees’ turnover intention. Finally, this study has limitations in the selection of the moderating variable. Conscientiousness was chosen as the moderating variable to explore its boundary conditions in the relationship between the uasge of robot and turnover intention. However, in the practical context of hotel management, some hotels may not assess employees’ personality traits during the recruitment process. This limitation may restrict the practical contribution of this study. Therefore, future research could consider alternative types of moderating variables, such as the intensity of human resource management, to ensure that the research findings have practical implications for management practices. Declaration of competing interest None. Acknowledgments This research was supported by the National Natural Science Foundation of China (72172032), the Natural Science Foundation of Hebei Province (G2021501006), and the National College Student Innovation and Entrepreneurship Training Program of China (CY2023002). Appendix Items measuring The usage of robots (Source: Medcof, 1996) 1 I used robots to carry out most of my job functions. 2 I spent most of the time working with robots. 3 I worked with robots in making major work decisions. Items measuring Work autonomy (Source: Spreitzer, 1995) 1 I have significant autonomy in determining how I do my job. 2 I can decide on my own how to go about doing my work. 3 I have considerable opportunity for independence and freedom in how I do my job. Items measuring Job insecurity (Source: Mauno and Ulla (2001) 1 Your job is insecure. 2 Your job is likely to change in the future. 3 Your job is not permanent. 4 You are worried about the possibility of being fired. 5 The thought of getting fired really scares you. Items measuring Turnover intention (Source: Mobley et al., 1979) 1. I basically have no desire to leave this current business. 2. I plan to have a long-term career in this business. 3. I often feel bored with my current job and want to change to a new business. 4. In the next six months, I will most likely leave this current business. Items measuring Conscientiousness (Source: Pathki et al., 2022) 1. At work, I get my tasks done right away. 2. I am careful to put things back in their proper place at work. 3. At work, I like order. 4. I am always prepared at work. References Alkahtani, A. H. (2015). Investigating factors that influence employees’ turnover intention: A review of existing empirical works. International Journal of Business and Management, 10(12), 152–166. Bamberger, P., & Belogolovsky, E. (2017). The dark side of transparency: How and when pay administration practices affect employee helping. Journal of Applied Psychology, 102(4), 658–671. Belete, A. (2018). Turnover intention influencing factors of employees: An empirical work review. Journal of Entrepreneurship & Organization Management, 7(3), 1–7. Bogg, T., & Roberts, B. W. (2004). Conscientiousness and health-related behaviors: A meta-analysis of the leading behavioral contributors to mortality. Psychological Bulletin, 130(6), 887–919. Bothma, C. F., & Roodt, G. (2013). The validation of the turnover intention scale. SA Journal of Human Resource Management, 11(1), 1–12. Brislin, R. W. (1970). Back-translation for cross-cultural research. Journal of CrossCultural Psychology, 1(3), 185–216. Budhwar, P., Malik, A., De Silva, M. T., & Thevisuthan, P. (2022). Artificial intelligence–challenges and opportunities for international HRM: A review and L.-X. Zhang et al.

Journal of Hospitality and Tourism Management 57 (2023) 74–83 83 research agenda. International Journal of Human Resource Management, 33(6), 1065–1097. Chang, H. Y., Huang, T. L., Wong, M. K., Ho, L. H., Wu, C. N., & Teng, C. I. (2021). How robots help nurses focus on professional task engagement and reduce nurses’ turnover intention. Journal of Nursing Scholarship, 53(2), 237–245. Cheng, H., Jia, R., Li, D., & Li, H. (2019). The rise of robots in China. The Journal of Economic Perspectives, 33(2), 71–88. Cheung, G. W., & Lau, R. S. (2017). Accuracy of parameter estimates and confidence intervals in moderated mediation models: A comparison of regression and latent moderated structural equations. Organizational Research Methods, 20(4), 746–769. Choi, Y., Choi, M., Oh, M., & Kim, S. (2020). Service robots in hotels: Understanding the service quality perceptions of human-robot interaction. Journal of Hospitality Marketing & Management, 29(6), 613–635. Cho, Y. J., & Lewis, G. B. (2012). Turnover intention and turnover behavior: Implications for retaining federal employees. Review of Public Personnel Administration, 32(1), 4–23. Fan, H., Gao, W., & Han, B. (2022). How does (im) balanced acceptance of robots between customers and frontline employees affect hotels’ service quality? Computers in Human Behavior, 133, Article 107287. Fuentes-Moraleda, L., Diaz-Perez, P., Orea-Giner, A., Munoz-Mazon, A., & VillaceMolinero, T. (2020). Interaction between hotel service robots and humans: A hotelspecific service robot acceptance model (sRAM). Tourism Management Perspectives, 36, Article 100751. Fu, S., Zheng, X., & Wong, I. A. (2022). The perils of hotel technology: The robot usage resistance model. International Journal of Hospitality Management, 102, Article 103174. Galletta, M., Portoghese, I., & Battistelli, A. (2011). Intrinsic motivation, job autonomy and turnover intention in the Italian healthcare: The mediating role of affective commitment. Journal of Management Research, 3(2), 1–19. Gardell, B. (1977). Autonomy and participation at work. Human Relations, 30(6), 515–533. Greenhalgh, L., & Rosenblatt, Z. (1984). Job insecurity: Toward conceptual clarity. Academy of Management Review, 9(3), 438–448. Grijalva, E., & Harms, P. D. (2014). Narcissism: An integrative synthesis and dominance complementarity model. Academy of Management Perspectives, 28(2), 108–127. Heider, F. (1982). The psychology of interpersonal relations. New York: Wiley. Hobfoll, S. E. (1989). Conservation of resources. A new attempt at conceptualizing stress. American Psychologist, 44(3), 513–524. Hobfoll, S. E., Halbesleben, J., Neveu, J. P., & Westman, M. (2018). Conservation of resources in the organizational context: The reality of resources and their consequences. Annual Review of Organizational Psychology and Organizational Behavior, 5(1), 103–128. Hu, J., & Judge, T. A. (2017). Leader–team complementarity: Exploring the interactive effects of leader personality traits and team power distance values on team processes and performance. Journal of Applied Psychology, 102(6), 935–955. Jaiswal, D., & Dhar, R. L. (2017). Impact of human resources practices on employee creativity in the hotel industry: The impact of job autonomy. Journal of Human Resources in Hospitality & Tourism, 16(1), 1–21. Jordan, P. J., Ashkanasy, N. M., & Hartel, C. E. (2002). Emotional intelligence as a moderator of emotional and behavioral reactions to job insecurity. Academy of Management Review, 27(3), 361–372. Kaiser, H. F. (1974). An index of factorial simplicity. Psychometrika, 39(1), 31–36. Khaliq, A., Waqas, A., Nisar, Q. A., Haider, S., & Asghar, Z. (2022). Application of AI and robotics in hospitality sector: A resource gain and resource loss perspective. Technology in Society, 68, Article 101807. Kiesler, D. J. (1983). The 1982 interpersonal circle: A taxonomy for complementarity in human transactions. Psychological Review, 90(3), 185–204. Koo, B., Curtis, C., & Ryan, B. (2021). Examining the impact of artificial intelligence on hotel employees through job insecurity perspectives. International Journal of Hospitality Management, 95, Article 102763. Lee, S. H., & Jeong, D. Y. (2017). Job insecurity and turnover intention: Organizational commitment as mediator. Social Behavior and Personality: International Journal, 45(4), 529–536. Lee, M. K., Kusbit, D., Metsky, E., & Dabbish, L. (2015). Working with machines: The impact of algorithmic and data-driven management on human workers. In Proceedings of the 33rd annual ACM conference on human factors in computing systems (pp. 1603–1612). Lee, Y., Lee, S., & Kim, D.-Y. (2021). Exploring hotel guests’ perceptions of using robot assistants. Tourism Management Perspectives, 37, Article 100781. Li, J. J., Bonn, M. A., & Ye, B. H. (2019). Hotel employee’s artificial intelligence and robotics awareness and its impact on turnover intention: The moderating roles of perceived organizational support and competitive psychological climate. Tourism Management, 73, 172–181. Li, J. M., Wu, T. J., Wu, Y. J., & Goh, M. (2023). Systematic literature review of human–machine collaboration in organizations using bibliometric analysis. Management Decision. https://doi.org/10.1108/MD-09-2022-1183 Li, J. M., Zhang, X. F., Zhang, L. X., & Zhang, R. X. (2023). Customer incivility and emotional labor: The mediating role of dualistic work passion and the moderating role of conscientiousness. Current Psychology. https://doi.org/10.1007/s12144-022- 04107-6 Luo, J. M., Vu, H. Q., Li, G., & Law, R. (2021). Understanding service attributes of robot hotels: A sentiment analysis of customer online reviews. International Journal of Hospitality Management, 98, Article 103032. Lu, V. N., Wirtz, J., Kunz, W. H., Paluch, S., Gruber, T., Martins, A., et al. (2020). Service robots, customers and service employees: What can we learn from the academic literature and where are the gaps? Journal of Service Theory Practice, 30(3), 361–391. Manolopoulos, D., Peitzika, E., Mamakou, X. J., & Myloni, B. (2022). Psychological and formal employment contracts, workplace attitudes and employees’ turnover intentions: Causal and boundary inferences in the hotel industry. Journal of Hospitality and Tourism Management, 51, 289–302. Mauno, S., & Ulla, K. L. (2001). Multi-wave, multi-variable models of job insecurity: Applying different scales in studying the stability of job insecurity. Journal of Organizational Behavior, 22(8), 919–937. Medcof, J. W. (1996). The job characteristics of computing and non-computing work activities. Journal of Occupational and Organizational Psychology, 69(2), 199–212. Menter, J. W. (1973). The chemical and petrochemical industries: Discussion. Philosophical Transactions of the Royal Society of London,Series A, 275(1250), 356-356. Mobley, W. H., Griffeth, R. W., Hand, H. H., & Meglino, B. M. (1979). Review and conceptual analysis of the employee turnover process. Psychological Bulletin, 86(3), 493–522. Nam, T. (2019). Technology usage, expected job sustainability, and perceived job insecurity. Technological Forecasting and Social Change, 138, 155–165. Osawa, H., Ema, A., Hattori, H., Akiya, N., Kanzaki, N., Kubo, A., … Ichise, R. (2017). What is real risk and benefit on work with robots? From the analysis of a robot hotel. In Proceedings of the companion of the 2017 ACM/IEEE international conference on human-robot interaction. Paper presented at the. Pathki, C. S., Kluemper, D. H., Meuser, J. D., & Mclarty, B. D. (2022). The Org-B5: Development of a short work frame-of-reference measure of the Big Five. Journal of Management, 48(5), 1299–1337. Podsakoff, P. M., MacKenzie, S. B., & Podsakoff, N. P. (2012). Sources of method bias in social science research and recommendations on how to control it. Annual Review of Psychology, 63(1), 539–569. Podsakoff, P. M., & Organ, D. W. (2016). Self-reports in organizational research: Problems and prospects. Journal of Management, 12(4), 531–544. Popa, I., Lee, L., Yu, H., & Madera, J. M. (2023). Losing talent due to COVID-19: The roles of anger and fear on industry turnover intentions. Journal of Hospitality and Tourism Management, 54, 119–127. Pu, B., Ji, S., & Sang, W. (2022). Effects of customer incivility on turnover intention in China’s hotel employees: A chain mediating model. Journal of Hospitality and Tourism Management, 50, 327–336. Schmidt-Wilk, J., Alexander, C. N., & Swanson, G. C. (1996). Developing consciousness in organizations: The transcendental meditation program in business. Journal of Business and Psychology, 10(4), 429–444. Song, Y., Zhang, M., Hu, J., & Cao, X. (2022). Dancing with service robots: The impacts of employee-robot collaboration on hotel employees’ job crafting. International Journal of Hospitality Management, 103, Article 103220. Spreitzer, G. M. (1995). Psychological empowerment in the workplace: Dimensions, measurement, and validation. Academy of Management Journal, 38(5), 1442–1465. Tang, P. M., Koopman, J., McClean, S. T., Zhang, J. H., Li, C. H., De Cremer, D., … Ng, C. T. S. (2021). When conscientious employees meet intelligent machines: An integrative approach inspired by complementarity theory and role theory. Academy of Management Journal, 65(3), 1019–1054. Vatan, A., & Dogan, S. (2021). What do hotel employees think about service robots? A qualitative study in Turkey. Tourism Management Perspectives, 37, Article 100775. Vrontis, D., Christofi, M., Pereira, V., Tarba, S., Makrides, A., & Trichina, E. (2022). Artificial intelligence, robotics, advanced technologies and human resource management: A systematic review. International Journal of Human Resource Management, 33(6), 1237–1266. Wang, L. H., Ho, J. L., Yeh, S. S., & Huan, T. C. T. (2022). Is robot hotel a future trend? Exploring the incentives, barriers and customers’ purchase intention for robot hotel stays. Tourism Management Perspectives, 43, Article 100984. Witte, H. D. (1999). Job insecurity and psychological well-being: Review of the literature and exploration of some unresolved issues. European Journal of Work & Organizational Psychology, 8(2), 155–177. Wu, T. J., Li, J. M., Wang, Y. S., & Zhang, R. X. (2023). The dualistic model of passion and the service quality of five-star hotel employees during the COVID-19 pandemic. International Journal of Hospitality Management, Article 103519. Wu, T. J., Li, J. M., & Wu, Y. J. (2022). Employees’ job insecurity perception and unsafe behaviours in human–machine collaboration. Management Decision, 60(9), 2409–2432. Wu, T. J., Zhang, R. X., & Li, J. M. (2023). How does emotional labor influence restaurant employees’ service quality during COVID-19? The roles of work fatigue and supervisor–subordinate guanxi. International Journal of Contemporary Hospitality Management. https://doi.org/10.1108/IJCHM-09-2022-1060 Xu, J., Zhu, D., & Li, Y. (2022). Does small and medium enterprise differential leadership increase subordinate knowledge hiding? Evidences from job insecurity, territorial consciousness and leadership performance expectation. Frontiers in Psychology, 13, Article 983669. Yam, K. C., Bigman, Y. E., Tang, P. M., Ilies, R., De Cremer, D., Soh, H., et al. (2021). Robots at work: People prefer—and forgive—service robots with perceived feelings. Journal of Applied Psychology, 106(10), 1557–1572. Yin, J., Bi, Y., & Ni, Y. (2022). The impact of COVID-19 on turnover intention among hotel employees: A moderated mediation model. Journal of Hospitality and Tourism Management, 51, 539–549. Yu, H., Shum, C., Alcorn, M., Sun, J., & He, Z. (2022). Robots can’t take my job: Antecedents and outcomes of Gen Z employees’ service robot risk awareness. International Journal of Contemporary Hospitality Management, 34(8), 2971–2988. Zhang, S., Hu, Z., Li, X., & Ren, A. (2022). The impact of service principal (service robot vs. human staff) on service quality: The mediating role of service principal attribute. Journal of Hospitality and Tourism Management, 52, 170–183. L.-X. Zhang et al.

Journal of Hospitality and Tourism Management 57 (2023) 133–142 Available online 22 September 2023 1447-6770/© 2023 The Authors. Published by Elsevier Ltd. on behalf of CAUTHE - COUNCIL FOR AUSTRALASIAN TOURISM AND HOSPITALITY EDUCATION. All rights reserved. How do I remind you? The combined effect of purchase motivation and reminding message content on tourism consumers’ verification behavior Mengmeng Song, Yuchen Wang * , Rui Guo School of Tourism, Hainan University, Haikou, China ARTICLE INFO Keywords: Tourism live-streaming Purchase motivation Reminding message Mental simulation Verification behavior ABSTRACT Field verification of tourism live-streaming products is related to the ultimate conversion of tourism benefits. Therefore, improving the product verification rate is critical for tourism companies focusing on live-stream marketing. Considering the targeted nature of purchasing tourism live-streaming products, a mechanism model is constructed to form consumer verification behavior based on goal-directed behavior theory, analyzing it through two scenario experiments. When tourism consumers purchase tourism products for promotional purposes, a near-expired discount message can better stimulate their verification behavior; when tourism consumers purchase tourism products for emotional purposes, a limited-time activity message can better stimulate their verification behavior. "Fear of missing out" (FOMO) partially mediated the above interactive relationship, while mental simulation played a moderating role in the primary effect relationship. This study’s results emphasize the mechanism of individual decision-making behavior after tourism live-streaming, providing a reference for the online marketing process for tourism companies. 1. Introduction The development of modern technology has provided new opportunities for the digital transformation of the tourism industry. An increasing number of tourism companies actively engage in livestreaming to promote tourism products and destinations (Deng et al., 2021). Tourism live-streaming not only breaks the boundaries of time and space (Delic et al., 2018) but also provides an immersive, entertaining, and authentic experience for tourism consumers through host interaction and live-streaming technology (Liu et al., 2022; Zheng et al., 2023). This gives them powerful capabilities to sell tourism products and recommend destinations (Xie et al., 2022; Zhang & Xiao, 2023). Therefore, the formation of consumers’ purchasing decision-making behavior in tourism live-streaming has received extensive attention. Scholars have focused on the influence of host characteristics (e.g., interactivity) or live-streaming perception (e.g., informativity, entertainment, presence) (Lin et al., 2022; Lv et al., 2022), but the verification behavior after the purchase of tourism live-streaming has rarely been studied. The verification behavior refers to the process of going to the field to confirm and verify the consumer’s booked tourism product (Inman & McAlister, 1994; Lv et al., 2020; Wang et al., 2015). This also characterizes how the tourism live-streaming product is different from general commodities, which can only be used as the power to use the product in the field at the agreed time period (Lv et al., 2022). For tourism enterprises, the realization of verification is related to the transformation and improvement of expected returns. However, due to the flexibility in the verification process of tourism live-streaming products, the current verification rate is low. According to Ctrip’s annual report data, in 2021, the verification rate of presale products in official Ctrip live-streaming rooms was only approximately 30%. This means that over 60% of the tourism products were either actively returned by consumers or automatically refunded because they exceeded the usage period. Therefore, it is insufficient for research to focus only on various marketing methods adopted in the process of tourism live-streaming (Song et al., 2021; Xie et al., 2022). It is also necessary to pay attention to the marketing measures from the completion of the purchase to the final field consumption stage. Based on the above research gaps and practical needs, this study focuses on effectively stimulating tourism consumers’ verification behavior. Unlike OTA products, tourism live-streaming products mainly adopt a “first purchase, then booking and verifying” model. In other words, consumers often do not need to choose a specific date when making a purchase for scenic spot tickets or hotel products. They need to select a date to make an appointment within a limited time (e.g., within 6 * Corresponding author. E-mail address: [email protected] (Y. Wang). Contents lists available at ScienceDirect Journal of Hospitality and Tourism Management journal homepage: www.elsevier.com/locate/jhtm https://doi.org/10.1016/j.jhtm.2023.09.009 Received 24 July 2023; Received in revised form 11 September 2023; Accepted 16 September 2023

Journal of Hospitality and Tourism Management 57 (2023) 133–142 134 months or on a holiday) after purchasing, and then go to the field for verification. They possess a certain degree of flexibility. As long as there is no appointment or verification, they can get a full refund, except for some package tour products (Lv et al., 2022). This creates a problem where consumers often do not immediately verify, and may even actively or passively give up verifying due to the passage of time. Therefore, to facilitate the ultimate conversion of tourism live-streaming revenue, tourism enterprises often use various forms of reminding messages (e.g., emails, text messages, and platform notifications) with different content to rekindle the expectations of tourists and motivate them to verify their products on-site. However, how reminding messages specifically affect consumer verification behavior as well as their psychological mechanism has not been addressed. Goal-directed behavior theory suggests that individual-specific behaviors are influenced by external and internal needs (Kim et al., 2021). Following this logic, reminding messages as an external incentive can only stimulate verification behavior when they align with the psychological needs of tourists (Hill et al., 2016), which typically manifest as purchase goals and motivations. Although consumers’ purchase motivations differ, they believe it will benefit them (Diaconu, 2015). Therefore, when the reminding messages create a sense of impending loss, they trigger a "fear of missing out" (FOMO), which in turn influences subsequent behavioral decisions (Good & Hyman, 2020). Additionally, the intangibility of tourism live-streaming products leads tourists to engage in different forms of mental simulation before verification, including process and outcome simulations (Taylor et al., 1998). Through mental simulation, individuals imagine possible future events to assess whether they should take corresponding actions (Lu & Jen, 2016). Mental simulation is often based on tourists’ previous travel experiences and serves their consumption goals; however, it also relies on various media such as videos, images, and textual information (Lv et al., 2020; Wang et al., 2023). For example, when faced with hotel information, consumers driven by leisure entertainment engage in process simulation (imagining the comfort of the bed), which enhances their willingness to make a reservation compared to outcome simulation (imagining good sleep quality) (Lv et al., 2020). Therefore, there may be a matching effect between consumer goals, information cues, and types of mental simulations on consumer behavior. Furthermore, goal-directed behavior theory suggests that individuals engage in mental simulations of goal attainment or failure to make behavioral decisions (Perugini & Bagozzi, 2001), and these goals typically manifest as travel consumer motivations (Ho et al., 2022). In other words, in the post-purchase phase, when consumers receive reminding messages, they will still conduct different forms of mental simulation based on their own needs to determine whether they want to verify. Previous studies have ignored this aspect, focusing only on mental simulation during the purchase process (Lv et al., 2020; Wang et al., 2023). Based on the above discussion, this study intends to answer the following three questions through the situational experiment method based on goal-oriented behavior theory: 1) How can the purchase motivation of tourism live-streaming products be effectively matched with the content of reminding messages to promote consumer verification behavior?; 2) What role does FOMO play?; and 3) What is the role of different forms of mental simulation in the above interaction? The findings of this study provide a fresh perspective for live-streaming marketing research and offer insights and references for tourism businesses to formulate post-purchase marketing strategies. 2. Literature review and research hypotheses 2.1. Goal-directed behavior theory Goal-directed behavior theory is a theoretical framework that considers individual behavioral motivations and emotional components to understand the formation of individual behaviors. It is recognized as advantageous compared to rational action theory and planned behavior theory and has been widely applied to explain goal-oriented decisionmaking (Lee et al., 2020; Perugini & Bagozzi, 2004). Goal-directed behavior theory primarily involves three aspects. First, individual behavior is influenced by external incentives in the environment and internal psychological needs (Moscarello et al., 2010), which often manifest as desires and motivations (Chiu & Cho, 2022). In this study, the motivation to purchase tourism live-streaming products and reminders correspond to internal psychological needs and external incentives, respectively, and jointly influence verification behavior. Second, individuals consider the situation of attainment or failure to achieve goals before making behavioral decisions and are guided by positive or negative emotional reactions in the decision-making process (Fry et al., 2014), such as anticipated regret and excitement. In this study, mental simulation corresponded to the anticipation of goal achievement, whereas the FOMO corresponded to an individual’s emotional response. Third, this theory emphasizes the direct impact of individuals’ past behavioral decisions on their current goal-directed behaviors (Han, 2021). For example, past purchasing behaviors of tourism live-streaming products directly drive future verification behaviors. Considering that decision-making in tourism live-streaming is similar to other forms of online shopping and falls under goal-directed behavior, this study intends to use goal-directed behavior theory as the theoretical foundation for the overall model to further expand its application boundaries. 2.2. Interaction of purchase motivation and reminding messages of tourism live-streaming products on verification behavior Drawing on the concept of online travel purchase motivation (Meng & Choi, 2016), the purchase motivation of tourism live-streaming products refers to the internal driving force that encourages consumers to implement purchase activities in the tourism live-streaming room. In tourism live-streaming, tourists often have clear goals and a clear understanding of the destination they want to visit or the type of tourism product they need, leading to different purchasing motivations (Deng et al., 2021). Current motivations for purchasing tourism live-streaming can be divided into promotional and emotion-oriented categories. The promotion-oriented category refers to tourists seeking unique price discounts or differentiated freebies, such as coupons, vouchers, cashback, meals, and souvenirs (Xie et al., 2022). The emotion-oriented category refers to tourists driven by interest, novelty, and other emotional factors, expecting emotional satisfaction by exploring and experiencing limited activities at tourist destinations (Lv et al., 2022). However, owing to the feature of full refunds for tourism live-streaming products anytime and anywhere (Xu et al., 2021), tourists are not inclined to visit the destination immediately for verification, resulting in an uncertain time gap between purchase and verification. Therefore, pushing reminding messages is important for tourism companies to promote verification behaviors. A reminding message is advertising information with a special purpose (Schwebel & Larimer, 2018), strengthening individuals’ consumption goals and effectively stimulating their consumption decisions (Moscarello et al., 2010). Research shows that reminding messages that match consumer goals can attract attention and increase intimacy and engagement, thereby increasing product sales in the short term or the future (Bies et al., 2021). Following this logic, for promotion-oriented tourists, near-expired discount messages can better align with their pursuit of discounts, thus promoting verification behavior. For emotion-oriented tourists, reminding messages about limited-time activities can better align with their pursuit of emotional enjoyment, enhancing their motivation for verification. This is because timely verification allows tourists to gain an additional entertainment experience and enhance their perceived value (Song et al., 2017). In addition, goal-directed behavior theory suggests that individual behavior is influenced by external stimuli, internal psychological needs, and past behaviors (Han, 2021; Moscarello et al., 2010). In this study, tourism M. Song et al.

Journal of Hospitality and Tourism Management 57 (2023) 133–142 135 purchasing motivations consider both internal psychological needs and the potential influence of past purchasing behavior on post-purchase verification behavior, with reminding messages serving as external stimuli. Therefore, this study posits that tourism purchasing motivations and reminding messages have an interactive impact on verification behavior. Based on the above, this study proposes the following hypotheses: H1. When tourists make purchasing decisions based on promotions, reminding them about near-expired discounts is more likely to stimulate their verification behavior. H2. When tourists make purchasing decisions based on emotions, reminding them about limited-time activities is more likely to stimulate their verification behavior. 2.3. Mediating effect of FOMO FOMO refers to a general concern, accompanied by a sense of consumption and anxiety; that is, fear that others may have beneficial experiences or better things that they do not have (Przybylski et al., 2013). In the field of tourism research, FOMO is manifested in the fear of missing opportunities such as leisure vacations, interactive social networking, and discounts (Zaman et al., 2022). FOMO often stems from the persuasive expression of businesses that can create an atmosphere of potential loss and transmit it to consumers (Good & Hyman, 2021). When tourists make purchasing decisions based on promotions, they indicate a desire for discounts to reduce their overall travel expenses. Reminding messages on near-expiration discounts can evoke anxiety and fear of losing the benefits they have obtained through their efforts (Gabler et al., 2017). One important reason for this phenomenon may be information asymmetry between tourists and travel companies, as tourists are often uncertain when they enjoy similar discounts (Ye et al., 2023). Consequently, they experience FOMO and make choices that maximize their utility, such as verifying promotions. In contrast, when tourists make purchase decisions based on emotions, they indicate their desire to relax and entertain themselves at their destination to achieve self-fulfillment. As tourists pursue immersive and heterogeneous experiences at their destinations, reminding them of time-limited activities is more effective in triggering their anticipation. In other words, they experience FOMO if they do not verify such activities in person (Zaman et al., 2022). This is because these types of tourists are more concerned about the kinds of experiences that the destination can offer them (Ryu et al., 2021). Similar viewpoints have been confirmed in tourism research, where individuals’ motivation for continuous learning matches the knowledge that museums can provide, leading to FOMO and ultimately positively influencing their intention to visit museums (Uslu & Tosun, 2023). Therefore, this study posits that when the purchasing motivation for live-streaming travel products aligns with reminding messages, it can better stimulate tourists’ FOMO. FOMO often leads consumers to engage in protective behaviors to avoid regretting inaction, such as making purchases or participating in activities (Good & Hyman, 2021). Previous research has shown that using advertisements containing FOMO messages for hedonic products can increase consumer purchase likelihood because consumers are more willing to pay for what they perceive as unmissable products or services (Dinh & Lee, 2022). Furthermore, from the definition of FOMO, we can see its influence on consumer decision-making. Individuals desire better travel products and higher-quality travel experiences than others, so they do not want to miss such opportunities (Abel et al., 2016). In summary, this study suggests that tourists’ FOMO often manifests as a reluctance to miss out on discounts offered by travel companies or limited activities at travel destinations and that this emotional response can lead to verification behaviors. Therefore, the following hypotheses is proposed: H3. When tourists make purchase decisions based on promotions, reminding messages about near-expired discounts are more likely to stimulate their FOMO, thereby leading to verification behaviors. H4. When tourists make purchase decisions based on emotions, reminding messages about limited-time activities are more likely to stimulate their FOMO, thereby leading to verification behaviors. 2.4. Moderating effect of mental simulation Mental simulation refers to the imitative psychological representation of real or imaginary events (Lu & Jen, 2016). When consumers are exposed to product information, even if they have never actually experienced it, they will spontaneously perform mental simulation based on their own experience or other similar consumption experience (Wu et al., 2021) and then show corresponding behavior according to their judgment (Royo-Vela & Black, 2020). This is not uncommon in tourism practice. For example, tourists can make mental simulations based on the text and picture clues provided by the hotel and then decide whether to book (Lv et al., 2020). Mental simulation can be divided into process and outcome (Taylor et al., 1998). In this study, process simulation emphasizes individuals’ imaginations of participating step-by-step in tourism activities, whereas outcome simulation emphasizes individuals’ imaginations of the ideal result of an event. Previous research has shown that process simulation has a greater influence on consumers with hedonic motives, whereas outcome simulation has a greater influence on consumers with utilitarian motives (Liu et al., 2022). Specifically, utilitarian motives correspond to purchase motivations based on promotions, whereas hedonic motives correspond to purchase motivations based on emotions. When the type of psychological simulation matches individuals’ goals, it can have a positive impact on their behavioral decisions, as motivation and goals are aligned (Zhao et al., 2011). Based on this logical approach, in a scenario in which tourism consumers make purchase decisions based on promotions, reminding messages about near-expiration discounts can match their motivations. Outcome simulations can prompt them to imagine themselves gaining advantages over others, thus enabling them to participate in tourism activities with greater benefits (Wang et al., 2023). Therefore, outcome simulation can better promote verification behavior compared to process simulation. In contrast, reminding messages about limited-time activities do not have a significant influence on them; instead, it may lead to the perception of additional costs and suspicion about the true motives of tourism companies, resulting in unintended effects (Schwerdtfeger et al., 2012). In this case, the process and outcome simulation had almost no influence on the verification behavior. In a scenario where tourism consumers make purchase decisions based on emotions, reminding messages about limited-time activities can match their motivations. Process simulation can evoke the imagination of joy and excitement brought about by participating in and experiencing unique activities at a tourism destination (Skard et al., 2021). Therefore, process simulation can better promote verification behavior compared to outcome simulation. Conversely, the reminding message about near-expiration discounts does not have a significant influence on them, as it does not align with the goals they want to achieve and may cause some level of aversion. In this case, the process and outcome simulation had almost no influence on the verification behavior. Based on the above, this study proposes the following hypotheses: H5. When tourism consumers make purchase decisions based on promotions and face reminding messages about near-expiration discounts, outcome simulation can more effectively stimulate their verification behavior. When reminding messages are about limited-time activities, there is no difference in the influence of process and outcome simulation on verification behavior. H6. When tourism consumers make purchase decisions based on M. Song et al.

Journal of Hospitality and Tourism Management 57 (2023) 133–142 136 emotions and face reminding messages about limited-time activities, process simulation can more effectively stimulate their verification behavior. When reminding messages about near-expired discounts are sent, there is no difference in the influence of process and outcome simulation on verification behavior. Based on this hypothetical derivation, the following theoretical model was constructed (Fig. 1). 3. Study 1: the direct and indirect effects of reminding messages on the verification behavior of tourism consumers under different purchase motivations 3.1. Research design The purpose of Study 1 was to test H1 to H4. Study 1 used 2 (purchase motivation for tourism live-streaming products: promotional motivation vs. emotional motivation) × 2 (reminding message: nearexpired discounts vs. limited-time activities) between-group design and selected tourist attraction ticket products as target products for consumers. The questionnaire comprised five sections, all used on a seven-point Likert scale. In each section, an attention test question was used to test whether the subjects filled out the questionnaire carefully; those who failed this test and those who filled out the questionnaire in less than 120 s were excluded. During the formal experiment, we would inform subjects about the form (text messages or platform notifications) of the reminding message and show the corresponding example. The first part involved the presentation of the stimulus material (Appendix A). This section includes the introduction and experimental materials related to these variables. The goal was to guide the subjects to quickly immerse themselves in a research scenario. We ensured that the word counts of the materials were as similar as possible to avoid a possible cognitive load on the experimental results. The second part, which measured subjects’ tourism product verification behavior, was mainly adapted from Lam and Hsu (2006) and Lv et al. (2022) and contained two questions, while FOMO was measured mainly based on Good and Hyman’s (2021) study and consisted of seven items (Appendix B). The third part involved manipulating the variables. For the control of purchase motivation, reference was mainly to the study of Liu et al. (2020) by scoring the presentation of information by subjects (fully promotion-based = 1, fully emotion-based = 7). For the control of reminding messages, reference was mainly to the study of Zhao et al. (2021) by scoring the presentation of information (near-expired discounts = 1, limited-time activities = 7). The fourth part is the measurement of distractors, that is, the control of the relevant variables. To avoid individuals reacting differently to the information material resulting from differences in preferences and thus biasing the results, this study measured the attractiveness of the information, mainly referring to Gramazio et al. (2021). Simultaneously, to avoid the effects of different individual responses on the truthfulness of the information, Study 1 also measured the credibility of the information, mainly referring to the study by Erkan and Evans (2016), which contains three items. The fifth section measured the subjects’ demographics, including gender, age, educational background, and discretionary monthly income. 3.2. Participants Study 1 primarily involved targeted research using the Wenjuanxing survey platform, inviting consumers who purchased live-streaming products to participate in the experiment. It is worth noting that Wenjuanxing is the earliest established company in the field of surveys in China and the most widely used and popular platform for questionnaire design and research, ensuring data confidentiality and effectiveness (Zhang et al., 2021). Before presenting the stimulus materials, subjects were asked to imagine a tourism attraction they had recently been interested in and then complete the corresponding questionnaire. Two hundred forty subjects were invited to participate in the study, and 223 valid questionnaires were collected, resulting in an effective response rate of 92.92%. The demographic characteristics of the subjects are as follows: Regarding gender, there were 108 males, accounting for 48.43%. Regarding age, the highest proportion was among those aged 18–30 (37.22%). Regarding educational background, the highest proportion (39.91%) was of subjects with a bachelor’s degree. Regarding monthly disposable income, the highest proportion was for those with a monthly income between 1001 and 3000 yuan, accounting for 33.18% of the total. 3.3. Results and analysis Regarding the reliability test of the variables, the results showed that Cronbach’s α of FOMO and verification behavior were 0.908 and 0.808, respectively, which were greater than 0.8, indicating that good reliability. For manipulation, the results showed that the type of purchase motivation for tourism live-steaming products was successfully manipulated, and there was a significant difference in subjects’ perceptions of the two purchase motivation materials (Mpromotional motivation = 2.000, SD = 0.838, Memotional motivation = 5.901, SD = 0.809, p < 0.001); the type of reminding message was successfully manipulated, and there was a significant difference in subjects’ perceptions of the reminding message materials (Mnear-expired discounts = 1.528, SD = 0.502, Mlimited-time activities = 6.452, SD = 0.517, p < 0.001). Next, the experiment tested the control variables. The results showed that the information attractiveness control was successful, and there was no significant difference in the perceived attractiveness of the subjects between the near-expired discounts message group and the limited-time activities message group (Mnear-expired discounts = 5.045, SD = 0.842, Mlimited-time activities = 4.982, SD = 0.726, p > 0.05); the information credibility control was successful, and the perceived credibility of the subjects between the near-expired discounts message group and the limited-time activities message group was not significant differences (Mnear-expired discounts = 6.036, SD = 0.793, Mlimited-time activities = 5.919, SD = 0.799, p > 0.05). Following this, a direct-effect analysis was performed. A two-factor Fig. 1. Hypothetical model. M. Song et al.

Journal of Hospitality and Tourism Management 57 (2023) 133–142 137 ANOVA was conducted using SPSS 26.0 software to test the interaction effect of purchase motivation and type of reminding messages on verification behavior, where tourism product verification behavior was used as the dependent variable, purchase motivation and reminding messages were used as fixed factors, and the test results are shown in Fig. 2. The results show that the interaction effect of the type of purchase motivation and type of reminding message on verification behavior for tourism live-streaming products was significant (F(1,219) = 979.420, p < 0.001). Specifically, for the promotion-based purchase motivation group, the effect of near-expired discounts messages on verification behavior was significantly higher than that of limited-time activities messages (Mnearexpired discounts = 5.418, SD = 0.504, Mlimited-time activities = 3.097, SD = 0.774, F(1,110) = 356.143, p < 0.001); for the emotion-based purchase motivation group, limited-time activities messages had a higher impact on verification behavior than near-expired discounts messages (Mnearexpired discounts = 3.226, SD = 0.562, Mlimited-time activities = 6.112, SD = 0.617, F(1,109) = 664.893, p < 0.001). In summary, H1 and H2 were supported. Finally, the mediating effects of the experiment were analyzed. First, a two-factor ANOVA was applied to test the interaction effect of purchase motivation and reminding message type on FOMO. The results showed a significant interaction effect (F(1,219) = 1627.492, p < 0.001). Specifically, for the promotion-based purchase motivation group, the near-expired discounts message elicited significantly higher FOMO than the limited-time activities message (Mnear-expired discounts = 4.740, SD = 0.412, Mlimited-time activities = 3.026, SD = 0.327, F(1,110) = 598.543, p < 0.001); for the emotion-based purchase motivation group, the limitedtime activities message elicited a higher FOMO than the near-expired discounts message (Mnear-expired discounts = 2.922, SD = 0.378, Mlimitedtime activities = 4.968, SD = 0.263, F(1,109) = 1111.598, p < 0.001). Second, a mediating effect analysis was conducted using Model 8 in the process plug-in embedded in the SPSS 26.0 software (Hayes, 2017) to examine the mediating effect of FOMO (where promotional motivation = 1; emotional motivation = 0; near-expired discounts = 1; limited-time activities = 0). The results showed that when the purchase motivation of tourism live-streaming products was promoted, and the subjects were given a reminding message about near-expiration discounts, the mediating effect of FOMO was significant (LLCI = 0.749, ULCI = 1.543, excluding 0), indicating the existence of the mediating effect, while the main effect was significant (LLCI = 0.755, ULCI = 1.643, excluding 0), indicating that FOMO played a partially mediating role; thus, H3 is supported. When the purchase motivation for tourism live-streaming products was emotional, and subjects were given a reminding message about limited-time activities, the mediating effect of FOMO was significant (LLCI = − 1.908, ULCI = − 0.883, excluding 0), while the main effect was significant (LLCI = − 1.996, ULCI = − 0.967, excluding 0), indicating that FOMO played a partially mediating role; thus, H4 was supported. Although Study 1 explored the influence of reminding messages on tourism consumers in the context of different purchase motivations and the mediating role played by FOMO, the possible role of individual mindsets has not yet been considered. In fact, when processing information, individuals rely on their own experiences and information cues to construct scenarios for using and experiencing the product in question (Yim et al., 2021), a mental simulation process. Given this, Study 2 focused on the critical role played by different types of mental simulations in the formation of verification behavior. 4. Study 2: the moderating effect of mental simulation 4.1. Research design The purpose of Study 2 was to test H5 to H6 and to re-test H1 and H2 in the replacement scenario. The experimental design of Study 2 was the same as that of Study 1, with the addition of the mental simulation type variable (process vs. outcome) to the original between-group design and the inclusion of mental simulation materials (process simulation = 1, outcome simulation = 0). Mental simulation manipulation usually uses textual or pictorial materials (Sanna et al., 1998). Study 2 selected textual materials based on live tourism practices, as shown in Appendix C, while other presentations and layouts were basically the same as in Study 1. The manipulation of mental simulation was mainly based on Escalas and Luce’s (2004) study, where subjects were scored by their perception of the information (full process simulation = 1, full outcome simulation = 7). 4.2. Participants Study 2 still used the Wenjuanxing research service and asked the subjects to imagine a hotel product they had recently paid attention to before the stimulus material was presented and then filled in the corresponding questionnaire. At the same time, Study 2 explained the meaning and difference between process and outcome simulation to the subjects to facilitate their understanding. Study 2 invited 320 subjects to participate, and 303 valid questionnaires were returned, with an effective rate of 94.69%. The personal statistical characteristics of the subjects were as follows: regarding gender, 47.85% were male, and 52.15% were female. Regarding age, 33.99% were between 18 and 30 years old. Regarding education level, 36.30% had a bachelor’s degree. Regarding disposable monthly income, the number of people in the range of 1001~3000 yuan, accounting for the highest percentage of 30.36%. 4.3. Results and analysis Regarding the reliability test of the variables, the results showed that Cronbach’s ɑ of the verification behavior was 0.803, which was greater than 0.8, indicating good reliability. For manipulation, the results showed that the type of purchase motivation to tourism live-streaming products was successfully manipulated, and there was a significant difference between subjects’ perceptions of the two purchase motivation materials (Mpromotional motivation = 1.605, SD = 0.547, Memotional motivation = 6.158, SD = 0.594, p < 0.001); the type of reminding message was successfully manipulated, and there was a significant difference between subjects’ perceptions of the reminding message materials (Mnear-expired discounts = 1.474, SD = 0.604, Mlimited-time activities = 6.486, SD = 0.651, p < 0.001); the mental model type was manipulated successfully, and there was a significant difference in subjects’ perceptions of the mental simulation material (Mprocess simulation = 1.579, SD = 0.599, Moutcome simulation = 6.361, SD = 0.593, p < 0.001). Next, the experiment was tested for control variables. The results Fig. 2. Effect of the interaction between purchase motivation and reminding messages on verification behavior. M. Song et al.

Journal of Hospitality and Tourism Management 57 (2023) 133–142 138 showed that there was no significant difference in the perceived attractiveness of the subjects between the near-expired discounts message group and the limited-time activities message group (Mnear-expired discounts = 5.000, SD = 0.735, Mlimited-time activities = 4.919, SD = 0.862, p > 0.05); the information credibility control was successful, and there was no significant difference in the perceived credibility of the subjects between the near-expired discounts message group and the limited-time activities message group (Mnear-expired discounts = 5.869, SD = 0.704, Mlimited-time activities = 6.108, SD = 0.875, p > 0.05). Following that, Study 2 conducted another test for the interaction effect of motivation to purchase tourism live-streaming products and reminding messages. The results showed that the interaction effect of the type of motivation to tourism live-streaming products and the type of reminding message on verification behavior was significant (F(1,299) = 163.071, p < 0.05). Specifically, specifically, for the promotion-based purchase motivation group, the effect of near-expired discounts messages on verification behavior was significantly higher than that of limited-time activities messages (Mnear-expired discounts = 5.196, SD = 0.827, Mlimited-time activities = 4.441, SD = 0.632, F(1,148) = 39.608, p < 0.001); for the emotion-based purchase motivation group, the effect of limited-time activities messages had a higher impact on verification behavior than near-expired discounts messages (Mnear-expired discounts = 4.390, SD = 0.616, Mlimited-time activities = 5.651, SD = 0.658, F(1,151) = 149.931, p < 0.001). In summary, H1 and H2 are once again supported. Finally, the moderating effect of the mental simulation was analyzed using a three-way ANOVA with verification behavior as the dependent variable. The results showed (Figs. 3 and 4) that the interaction effects of the type of purchase motivation for tourism live-streaming products, the type of reminding message, and the type of mental model on verification behavior were significant (F(1,295) = 4.423, p < 0.05), indicating that the interaction effect of the type of mental model on the type of purchase motivation for tourism live-streaming products and the type of reminding message had a moderating effect. Specifically, when subjects’ purchase motivation was promotional, the outcome simulation was more able to stimulate their verification behavior compared to the process simulation when faced with the near-expired discounts messages (Moutcome simulation = 5.833, SD = 0.431, Mprocess simulation = 4.592, SD = 0.635, p < 0.001); there was no significant difference in the effect of the two simulations on verification behavior when faced with the limitedtime activities messages (Moutcome simulation = 4.487, SD = 0.568, Mprocess simulation = 4.392, SD = 0.699, p > 0.05). When subjects’ purchase motivation was emotional and faced with limited-time activities messages, process simulation was more able to motivate their verification behavior compared to outcome simulation (Moutcome simulation = 5.250, SD = 0.476, Mprocess simulation = 6.053, SD = 0.567, p < 0.001); the effect of the two models on verification behavior was not significantly different when faced with the near-expired discounts messages (Moutcome simulation = 4.282, SD = 0.605, Mprocess simulation = 4.500, SD = 0.615, p > 0.05). In summary, H5 and H6 were both supported. 5. General discussion Our study is based on goal-directed behavior theory and analyzes the complex mechanisms of how purchasing motivation and reminding messages interact to influence tourism consumers’ verification behavior through two scenario experiments. The conclusions and discussions are as follows: First, purchasing motivation and reminding messages have an interactive effect on verification behavior. When tourism consumers purchase based on promotions, sending near-expiration discount messages can better stimulate their verification behavior. When tourism consumers purchase based on emotions, sending limited time-activity messages can stimulate their verification behavior. In other words, providing reminding messages that align with the purchasing goals of tourism consumers can lead to subsequent product stickiness behaviors (Bies et al., 2021). Compared with previous studies that mainly focused on the impact of reminding consumers to influence purchase decisions, this study further explored the influence of reminding consumers about post-purchase verification, focusing on message content (Li et al., 2021). We believe this study is an effective extension of the research on tourism online marketing. Second, FOMO mediates the interactive effect of purchasing motivation and reminding messages on verification behavior in livestreaming tourism products. Similar to previous studies, social media information can stimulate consumers’ emotional responses (Liu & Huang, 2023). Meanwhile, this study again confirms that FOMO, as an emotional response, can stimulate tourism consumer demand in the post-consumption stage (Hodkinson, 2019), leading to positive behavioral decisions. This differs from previous studies that only emphasized the role of FOMO in the pre-purchase or purchase process (Good & Hyman, 2021). However, the incentive effect of FOMO cannot be ignored at any purchase stage. One possible reason is that FOMO can often stimulate inaction regret in tourism consumers. They might believe it is difficult to receive compensation if they do not act quickly, thus showing protective behavior, such as purchase and verify (Good & Hyman, 2020). Third, the type of mental simulation moderates the interactive effect of purchasing motivation and reminding messages on verification behavior in live-streaming tourism products. The results of this study Fig. 3. Moderating effect of mental simulation (promotional motivation). demonstrate that mental simulation is a direct driving force behind Fig. 4. Moderating effect of mental simulation (emotional motivation). M. Song et al.

Journal of Hospitality and Tourism Management 57 (2023) 133–142 139 tourism consumers’ proactive decision-making behaviors (Skard et al., 2021). Previous studies have emphasized mental simulation based on graphic and video information in the consumption process (Lv et al., 2020; Wang et al., 2023). In contrast, the results of this study emphasize that in the post-purchase phase, consumers will still conduct mental simulation based on their experience and reminding messages. 5.1. Theoretical implications First, this study expands the research perspective on decision-making in tourism live-streaming. Previous research has mainly analyzed tourism live-streaming behavior decisions from two perspectives. The first is purchasing decision behavior, such as the social and physical presence stimulated by tourism live-streaming, which encourages consumers to purchase (Xu et al., 2021). The other is travel decision behavior, such as generating on-site travel intentions from destination image, interaction, and product presence (Zhang et al., 2021). An apparent limitation is that the existing research only focuses on the behavior changes of tourism consumers during the tourism live-streaming process (Yang et al., 2022), and there is limited discussion after tourism live-streaming, especially about the verification behavior. Considering that tourism live-streaming products have delayed consumption, consumers may experience psychological changes such as hesitation, forgetting, and product comparison, resulting in abandonment of verification. Therefore, it important to understand how to stimulate the verification behavior. Furthermore, the timely verification of tourism live-streaming products is related to the collection and analysis of the final sales data of enterprises, the reasonable procurement of resources, and the optimization and improvement of services. Meanwhile, studying verification behavior can help tourism scholars better understand consumers’ psychological changes after purchase. Therefore, the results of this study respond to the prospect proposed by Lin et al. (2022), who suggest researching live-streaming tourism from various perspectives. Second, this study discusses the formation of verification behavior from the information content perspective in goal-directed behavior theory. Reminding messages provide great convenience for tourism consumers, enabling them to obtain accurate and rapid access to tourism product information, enhancing their purchase rate and loyalty (˙ Ilhan & Çeltek, 2016). However, previous research has primarily focused on the influence of reminding messages on individual consumer behavioral decision-making using theories such as the cognitive dissonance theory, integrative information acceptance model, and flow experience theory (Tan & Ooi, 2018; Zhao et al., 2021). Few studies have considered matching the delivery of information content with consumer goals. Therefore, this study explores the interactive influence of reminding messages and consumer purchasing motivation on verification behavior from the perspective of goal-directed behavior theory, expanding the application of reminding messages in tourism theoretical research to a certain degree and responding to the proposal of Li et al. (2022) to extend reminding message-related research to the field of tourism e-commerce. Third, this study emphasizes the importance of FOMO and mental simulation in forming verification behavior, thus contributing to a deeper understanding of decision-making behavior in tourism livestreaming. On the one hand, FOMO has been chiefly studied in media research to explain the negative impacts of mobile phone and social media usage on individuals; little research integrates this concept into the formation of consumer behavior in tourism (Zaman et al., 2022). Fortunately, in recent years, scholars have begun to pay attention to applying FOMO in consumer behavior research, pointing out that FOMO generated in specific consumption scenarios can stimulate positive consumer behavior (Good & Hyman, 2021). Therefore, this study expands the application of FOMO in tourism marketing research to explain the generation of post-purchase behavior by tourists. It also responds to the emphasis of Azemi et al. (2022) on a comprehensive understanding of consumer perceptions of reminders. On the other hand, in previous tourism research, scholars have usually treated mental simulation as a whole without categorizing it as a boundary condition (Bogicevic et al., 2019; Xie et al., 2023). However, individual trait differences may lead to different preferences for mental simulations. For example, business travelers tend to be performance-oriented and prefer outcome simulation, whereas leisure travelers tend to be entertainment-oriented and prefer process simulation (Lv et al., 2020). Different types of simulations have different effects on behavior. Therefore, this study focused on the influence of different types of mental simulation on individual behavior in different purchasing motivation backgrounds, which helps enrich the application of mental simulation in tourism consumer behavior studies. 5.2. Practical implications First, tourism enterprises should pay attention to the richness of the content and form of reminding messages. Creative elements should be added to the reminding message so that it is not too formal; information should be relayed with exciting pictures, unique emojis, and animation effects to attract consumers’ attention to the verification process. Introducing a gamification mechanism within the information link enables consumers to obtain additional benefits by experiencing puzzles, treasure hunts, or role-playing, such as free upgrade rooms, special dishes, and coupons. Meanwhile, the game can emphasize public welfare attributes to enhance consumer stickiness. We should also pay attention to the construction of the atmosphere of FOMO, especially to create a variety of reward mechanisms, such as a lottery and exclusive offers and experiences. Second, tourism companies should consider mental simulation’s role in verifying live-streaming products and providing consumers with more reference clues. Clues such as images or short video links can be included as reminders, allowing consumers to view and anticipate their experiences. Additionally, integrating metaverse or mixed reality technologies can provide consumers with corresponding links in platform notifications, enabling them to preview destination scenes through virtual avatars or to calculate the additional benefits they can gain compared to others through simulation programs. Furthermore, it is possible to use virtual digital human guides to interact with them for further mental simulation. 5.3. Limitations and future studies Although this study achieved its goals, it had some limitations. First, this study only considered reminding messages to promote the verification of live tourism streaming products. In the future, other strategies, such as product substitution or upgrade options of equal value, could be integrated to verify their effectiveness further. Second, this study only selected Chinese consumers as subjects. Future research could conduct comparative studies among cross-national tourism consumers to consider the potential impact of sociocultural factors. Third, this study only analyzed and validated the influencing mechanism model using experimental methods. Future research could enhance the external validity of these findings by utilizing real platform data or conducting longitudinal surveys. Fourth, this study only discussed the positive impact of reminding messages on verification behavior. Future research could also discuss the potential negative effects, such as invasiveness and information overload. Funding This work was supported by < China Natural Science Foundation > No. [72062015]; <Hainan Provincial Natural Science Foundation Project of China > No. [620RC561, 623RC443]; Hainan Province Key R&D Program(ZDYF2022GXJS007, ZDYF2022GXJS010). M. Song et al.

Journal of Hospitality and Tourism Management 57 (2023) 133–142 140 Data availability statement The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation. Authors’ contributions Mengmeng Song directed the study, Yuchen Wang drafted the study and analyzed the data, Rui Guo gived some useful advice for this study. All authors gave final approval of the manuscript before submission. Declaration of competing interest The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Acknowledgments Thanks to PhD Student Yuchen Jiao, School of Management, Jinan University, for his valuable suggestions in the process of manuscript writing and revision. Thanks to the tourists who helped us fill out the questionnaire. Appendix A. Stimulus material of Study 1 Promotional motivation Emotional motivation Reminding messages of near-expired discounts Reminding messages of limited-time activities Scenario display Suppose you purchased tourism livestreaming products with the intention of getting more discounts to minimize your overall travel expenses. Suppose you purchased tourism live-streaming products because of your interest (or curiosity) in the tourist attraction. Dear friends, Hello! The ticket product you have purchased is expiring soon, and the discount is about to expire as well. For more details, please click the link × × . Dear friends, Hello! The tourist attraction will soon hold a festive celebration event with limited-time access to themed activities. For more details, please click the link × × . Introductory remarks Please take a moment to calm down and compose yourself to consider the following scenario: Recently, you purchased a ticket to a tourism attraction through a tourism live-streaming platform, but you haven’t verified the ticket in person yet. As a result, the business has sent you a reminding message via text messages (assuming there is no fraudulent activity involved). Please read the specific materials below carefully and fill out the questionnaire based on your actual feelings. Appendix B. Scale of variables Variable type Variable name Items control variable Information attraction According to the description of the information, how attractive is the product to you? Information credibility I think the information is credible. I think the information is accurate. I think the information is somewhat persuasive. latent variable fear of missing out I am afraid that I will regret not verifying in time. I am afraid that I will miss something. I am afraid that others will experience this special activity/offer more discounts than me. I am afraid that I will feel anxious about not verifying in time. I would feel anxious about I would feel annoyed that I missed the verification. I will feel sorry that I did not experience the featured activities/lost the discount. I will worry that there will not be such a featured activity/discount again in the near future. verification behavior I will go to the field to verify the travel product within the expiration date. I will go to the field to verify the travel product within a short period of time after receiving the reminding message. Appendix C. Stimulus material of Study 2 process simulation outcome simulation Scenario display Please take a moment to imagine the experience of purchasing and staying at a hotel or homestay product through tourism live-streaming. For example, imagine lying on a comfortable bed, experiencing the interior decor, and participating in activities such as guessing games. Please take a moment to imagine the outcomes of experiencing the hotel or homestay product that you purchased through tourism live-streaming. For example, imagine receiving a discounted price, enjoying an afternoon tea, receiving a complimentary breakfast, and receiving complimentary tickets to the scenic area. References Abel, J. P., Buff, C. L., & Burr, S. A. (2016). Social media and the fear of missing out: Scale development and assessment. Journal of Business & Economics Research, 14(1), 33–44. Azemi, Y., Ozuem, W., Wiid, R., & Hobson, A. (2022). Luxury fashion brand customers’ perceptions of mobile marketing: Evidence of multiple communications and marketing channels. Journal of Retailing and Consumer Services, 66, Article 102944. Bies, S. M. T. A., Bronnenberg, B. J., & Gijsbrechts, E. (2021). How push messaging impacts consumer spending and reward redemption in store-loyalty programs. International Journal of Research in Marketing, 38(4), 877–899. Bogicevic, V., Seo, S., Kandampully, J. A., Liu, S. Q., & Rudd, N. A. (2019). Virtual reality presence as a preamble of tourism experience: The role of mental imagery. Tourism Management, 74, 55–64. Chiu, W., & Cho, H. (2022). The model of goal-directed behavior in tourism and hospitality: A meta-analytic structural equation modeling approach. Journal of Travel Research, 61(3), 637–655. M. Song et al.

Journal of Hospitality and Tourism Management 57 (2023) 133–142 141 Delic, A., Neidhardt, J., Nguyen, T. N., & Ricci, F. (2018). An observational user study for group recommender systems in the tourism domain. Information Technology & Tourism, 19(1–4), 87–116. Deng, Z., Benckendorff, P., & Wang, J. (2021). Travel live streaming: An affordance perspective. Information Technology & Tourism, 23(2), 189–207. Diaconu, V. I. (2015). New trends in the motivation behind buying luxury textile products. International Journal of Economic Practices and Theories, 5(5), 455–461. Dinh, T. C. T., & Lee, Y. (2022). ‘I want to be as trendy as influencers’-how “fear of missing out” leads to buying intention for products endorsed by social media influencers. The Journal of Research in Indian Medicine, 16(3), 346–364. Erkan, I., & Evans, C. (2016). The influence of eWOM in social media on consumers’ purchase intentions: An extended approach to information adoption. Computers in Human Behavior, 61, 47–55. Escalas, J. E., & Luce, M. F. (2004). Understanding the effects of process-focused versus outcome-focused thought in response to advertising. Journal of Consumer Research, 31(2), 274–285. Fry, M. L., Drennan, J., Previte, J., White, A., & Tjondronegoro, D. (2014). The role of desire in understanding intentions to drink responsibly: An application of the model of goal-directed behaviour. Journal of Marketing Management, 30(5–6), 551–570. Gabler, C. B., Myles Landers, V. M., & Reynolds, K. E. (2017). Purchase decision regret: Negative consequences of the steadily increasing discount strategy. Journal of Business Research, 76, 201–208. Good, M. C., & Hyman, M. R. (2020). ‘Fear of missing out’: Antecedents and influence on purchase likelihood. Journal of Marketing Theory and Practice, 28(3), 330–341. Good, M. C., & Hyman, M. R. (2021). Direct and indirect effects of fear-of-missing-out appeals on purchase likelihood. Journal of Consumer Behaviour, 20(3), 564–576. Gramazio, S., Cadinu, M., Guizzo, F., & Carnaghi, A. (2021). Does sex really sell? Paradoxical effects of sexualization in advertising on product attractiveness and purchase intentions. Sex Roles, 84(11–12), 701–719. Han, H. (2021). Consumer behavior and environmental sustainability in tourism and hospitality: A review of theories, concepts, and latest research. Journal of Sustainable Tourism, 29(7), 1021–1042. Hayes, A. F. (2017). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. New York: Guilford Publications. Hill, K. M., Fombelle, P. W., & Sirianni, N. J. (2016). Shopping under the influence of curiosity: How retailers use mystery to drive purchase motivation. Journal of Business Research, 69(3), 1028–1034. Ho, J. L., Chen, K. Y., Wang, L. H., Yeh, S. S., & Huan, T. C. (2022). Exploring the impact of social media platform image on hotel customers’ visit intention. International Journal of Contemporary Hospitality Management, 34(11), 4206–4226. Hodkinson, C. (2019). ‘Fear of missing out’(FOMO) marketing appeals: A conceptual model. Journal of Marketing Communications, 25(1), 65–88. ˙ Ilhan, ˙ I., & Çeltek, E. (2016). Mobile marketing: Usage of augmented reality in tourism. Gaziantep University Journal of Social Sciences, 15(2), 581–599. Inman, J. J., & McAlister, L. (1994). Do coupon expiration dates affect consumer behavior? Journal of Marketing Research, 31(3), 423–428. Kim, J. S., Lee, T. J., & Kim, N. J. (2021). What motivates people to visit an unknown tourist destination? Applying an extended model of goal-directed behavior. International Journal of Tourism Research, 23(1), 13–25. Lam, T., & Hsu, C. H. C. (2006). Predicting behavioral intention of choosing a travel destination. Tourism Management, 27(4), 589–599. Lee, C. K., Ahmad, M. S., Petrick, J. F., Park, Y. N., Park, E., & Kang, C. W. (2020). The roles of cultural worldview and authenticity in tourists’ decision-making process in a heritage tourism destination using a model of goal-directed behavior. Journal of Destination Marketing & Management, 18, Article 100500. Li, L., Li, X., Qi, W., Zhang, Y., & Yang, W. (2022). Targeted reminders of electronic coupons: Using predictive analytics to facilitate coupon marketing. Electronic Commerce Research, 22(2), 321–350. Li, J., Luo, X., Lu, X., & Moriguchi, T. (2021). The double-edged effects of e-commerce cart retargeting: Does retargeting too early backfire? Journal of Marketing, 85(4), 123–140. Lin, K., Fong, L. H. N., & Law, R. (2022). Live streaming in tourism and hospitality: A literature review. Asia Pacific Journal of Tourism Research, 27(3), 290–304. Liu, H., Feng, S., & Hu, X. S. (2022). Process vs. outcome: Effects of food photo types in online restaurant reviews on consumers’ purchase intention. International Journal of Hospitality Management, 102, Article 103179. Liu, C., & Huang, X. (2023). Does the selection of virtual reality video matter? A laboratory experimental study of the influences of arousal. Journal of Hospitality and Tourism Management, 54, 152–165. Liu, Z., Lei, S. H., Guo, Y. L., & Zhou, Z. A. (2020). The interaction effect of online review language style and product type on consumers’ purchase intentions. Palgrave Communications, 6(1), 1–8. Lu, M., & Jen, W. (2016). Effects of product option framing and temporal distance on consumer choice: The moderating role of process versus outcome mental simulations. Psychology and Marketing, 33(10), 856–863. Lv, X., Li, H., & Xia, L. (2020). Effects of haptic cues on consumers’ online hotel booking decisions: The mediating role of mental imagery. Tourism Management, 77, Article 104025. Lv, X., Zhang, R., Su, Y., & Yang, Y. (2022). Exploring how live streaming affects immediate buying behavior and continuous watching intention: A multigroup analysis. Journal of Travel & Tourism Marketing, 39(1), 109–135. Meng, B., & Choi, K. (2016). The role of authenticity in forming slow tourists’ intentions: Developing an extended model of goal-directed behavior. Tourism Management, 57, 397–410. Moscarello, J. M., Ben-Shahar, O., & Ettenberg, A. (2010). External incentives and internal states guide goal-directed behavior via the differential recruitment of the nucleus accumbens and the medial prefrontal cortex. Neuroscience, 170(2), 468–477. Perugini, M., & Bagozzi, R. P. (2001). The role of desires and anticipated emotions in goal-directed behaviours: Broadening and deepening the theory of planned behaviour. British Journal of Social Psychology, 40(1), 79–98. Perugini, M., & Bagozzi, R. P. (2004). The distinction between desires and intentions. European Journal of Social Psychology, 34(1), 69–84. Przybylski, A. K., Murayama, K., Dehaan, C. R., & Gladwell, V. (2013). Motivational, emotional, and behavioral correlates of fear of missing out. Computers in Human Behavior, 29(4), 1841–1848. Royo-Vela, M., & Black, M. (2020). Drone images versus terrain images in advertisements: Images’ verticality effects and the mediating role of mental simulation on attitude towards the advertisement. Journal of Marketing Communications, 26(1), 21–39. Ryu, S., Choi, K., & Cho, D. (2021). A behaviour-based typology of travellers using an online travel marketplace. Current Issues in Tourism, 24(2), 228–246. Sanna, L. J., Meier, S., & Turley-Ames, K. J. (1998). Mood, self-esteem, and counterfactuals: Externally attributed moods limit self-enhancement strategies. Social Cognition, 16(2), 267–286. Schwebel, F. J., & Larimer, M. E. (2018). Using text message reminders in health care services: A narrative literature review. Internet Interventions, 13, 82–104. Schwerdtfeger, A. R., Schmitz, C., & Warken, M. (2012). Using text messages to bridge the intention-behavior gap? A pilot study on the use of text message reminders to increase objectively assessed physical activity in daily life. Frontiers in Psychology, 3, 270. Skard, S., Knudsen, E. S., Sjåstad, H., & Thorbjørnsen, H. (2021). How virtual reality influences travel intentions: The role of mental imagery and happiness forecasting. Tourism Management, 87, Article 104360. Song, M., Choi, S., & Moon, J. (2021). Limited time or limited quantity? The impact of other consumer existence and perceived competition on the scarcity messaging–purchase intention relation. Journal of Hospitality and Tourism Management, 47, 167–175. Song, T. H., Kim, S. Y., & Ko, W. L. (2017). Developing an effective loyalty program using goal-gradient behavior in tourism industry. Journal of Travel & Tourism Marketing, 34 (1), 70–81. Tan, G. W. H., & Ooi, K. B. (2018). Gender and age: Do they really moderate mobile tourism shopping behavior? Telematics and Informatics, 35(6), 1617–1642. Taylor, S. E., Pham, L. B., Rivkin, I. D., & Armor, D. A. (1998). Harnessing the imagination: Mental simulation, self-regulation, and coping. American Psychologist, 53(4), 429–439. Uslu, A., & Tosun, P. (2023). Examining the impact of the fear of missing out on museum visit intentions. Journal of Hospitality & Tourism Research, Article 10963480231168608. Wang, X., Lai, I. K. W., Lu, Y., & Liu, X. (2023). Narrative or non-narrative? The effects of short video content structure on mental simulation and resort brand attitude. Journal of Hospitality Marketing & Management, 32(5), 593–614. Wang, X., Zhang, J. H., & Wu, X. G. (2015). Determinants of tourism coupon redemption. Journal of Travel & Tourism Marketing, 32(4), 339–351. Wu, L. L., Liu, S. Q., Huang, H., & Yu, X. (2021). Photo vs. art? The design of consumption guidance in cultural food consumption. International Journal of Hospitality Management, 97, Article 103008. Xie, C., Yu, J., Huang, S. S., & Zhang, J. (2022). Tourism e-commerce live streaming: Identifying and testing a value-based marketing framework from the live streamer perspective. Tourism Management, 91, Article 104513. Xie, Z., Zhang, M., & Ma, Z. (2023). The impact of mental simulation on subsequent tourist experience–dual evidence from eye tracking and self-reported measurement. Current Issues in Tourism, 26(18), 2915–2930. Xu, X., Huang, D., & Shang, X. (2021). Social presence or physical presence? Determinants of purchasing behaviour in tourism live-streamed shopping. Tourism Management Perspectives, 40, Article 100917. Yang, J., Zeng, Y., Liu, X., & Li, Z. (2022). Nudging interactive cocreation behaviors in live-streaming travel commerce: The visualization of real-time danmaku. Journal of Hospitality and Tourism Management, 52, 184–197. Ye, X., Fu, Y. K., Wang, H., & Zhou, J. (2023). Information asymmetry evaluation in hotel e-commerce market: Dynamics and pricing strategy under pandemic. Information Processing & Management, 60(1), Article 103117. Yim, M. Y. C., Kim, Y. K., & Lee, J. (2021). How to easily facilitate consumers’ mental simulation through advertising: The effectiveness of self-referencing image dynamics on purchase intention. International Journal of Advertising, 40(5), 810–834. Zaman, U., Barnes, S. J., Abbasi, S., Anjam, M., Aktan, M., & Khwaja, M. G. (2022). The bridge at the end of the world: Linking Expat’s pandemic fatigue, travel FOMO, destination crisis marketing, and vaxication for “greatest of all trips”. Sustainability, 14(4), 2312. Zhang, W., Wang, Y., & Zhang, T. (2021). Can ‘live-streaming’ really drive visitors to the destination? From the aspect of ‘social presence’. Sage open, 11(1), Article 21582440211006691. Zhang, A., & Xiao, H. (2023). Psychological well-being in tourism live streaming: A grounded theory. Journal of Hospitality & Tourism Research, 10963480221149595. M. Song et al.

Journal of Hospitality and Tourism Management 57 (2023) 133–142 142 Zhao, M., Hoeffler, S., & Zauberman, G. (2011). Mental simulation and product evaluation: The affective and cognitive dimensions of process versus outcome simulation. Journal of Marketing Research, 48(5), 827–839. Zhao, H., Wang, X., & Jiang, L. (2021). To purchase or to remove? Online shopping cart warning pop-up messages can polarize liking and purchase intention. Journal of Business Research, 132, 813–836. Zheng, S., Wu, M., & Liao, J. (2023). The impact of destination live streaming on viewers’ travel intention. Current Issues in Tourism, 26(2), 184–198. M. Song et al.

Journal of Hospitality and Tourism Management 57 (2023) 170–178 Available online 11 October 2023 1447-6770/© 2023 The Authors. Published by Elsevier Ltd. on behalf of CAUTHE - COUNCIL FOR AUSTRALASIAN TOURISM AND HOSPITALITY EDUCATION. All rights reserved. Influence of livestreamers’ intimate self-disclosure on tourist responses: The lens of parasocial interaction theory Yan Lu a,b , Xinyu Liu c,* , Yue Hu b , Chris Zhu d a School of Hospitality Administration, Zhejiang Yuexiu University, Shaoxing, Zhejiang, China b Institute for Research on Portuguese-Speaking Countries, City University of Macau, Taipa, Macau SAR, China c Faculty of International Tourism and Management, City University of Macau, Taipa, Macau SAR, China d School of Tourism Management, Macao Institute for Tourism Studies, Macau, China ARTICLE INFO Keywords: Livestreaming Intimate self-disclosure Parasocial interaction Parasocial relationship Purchase intention ABSTRACT Livestreaming has become a new channel for tourism marketing due to its high real-time and interactive characteristics. However, there is not a strong theoretical underpinning in the literature now available to explain parasocial phenomena in livestreaming. This study intends to fill this knowledge gap by expanding our knowledge of two different types of parasocial phenomena (parasocial interaction and parasocial relationship), how they are formed, and how they affect the audience. Therefore, this study proposes a theoretical framework that includes livestreamers’ intimate self-disclosure, parasocial interaction, parasocial relationships, trustworthiness, and purchase intention. The findings demonstrate that the intimate self-disclosure of travel livestreams enhances the parasocial relationship and parasocial interaction between audiences and livestreams and leads to the generation of audiences’ credibility, thus increasing their purchase intention. In addition, the research findings have important practical implications for tourism marketing, livestreaming platforms, and tourism livestreaming organizations. 1. Introduction With the rapid growth of information and communication technology, livestreaming has become popular worldwide and has been quickly adopted by the tourism industry (Xie et al., 2022). Travel livestreaming can minimize information interaction delays and provide more immediacy than traditional online tourism businesses (Yang, Zeng, et al., 2022). Therefore, travel livestreaming has become a new channel for destination marketing agencies to sell tourism products and promote destinations (Zheng et al., 2023). However, research on travel livestreaming is still in its infancy and requires a sound theoretical foundation to explain and forecast audiences’ psychological and behavioral intentions (Deng et al., 2022). To address the gaps in the current theoretical foundation for travel livestreaming research, we use parasocial interaction theory to construct and guide this research. Parasocial interaction theory is used to explain imagined social relationships and interactions between audiences and media characters (Rasmussen, 2018). At present, the application of parasocial interaction theory in marketing is mainly carried out from the parasocial relationship and parasocial interaction perspective to study the audience’s psychological and behavioral intentions in different technical environments (Chen et al., 2022; Lueck, 2015; Yang, Zhang, et al., 2022). Different from traditional media, in the context of travel livestreaming, the parasocial phenomenon (parasocial interaction and parasocial relationship) is no longer just a one-way imaginary relationship between media influencers and audiences (Deng et al., 2022). It is the driving force behind promoting the positive interaction and development of the relationship between media influencers and audiences (Tsay-Vogel & Schwartz, 2014). However, empirical research on parasocial phenomena in the live tourism environment is still in its early stages. In addition, many scholars often confuse parasocial interaction and parasocial relationship because Horton and Wohl (1956) did not discriminate between them when they first introduced them. Recent research has indicated that these two parasocial phenomena, while interrelated, are distinct concepts (Dibble et al., 2016; McLaughlin & Wohn, 2021) and may have different psychological and behavioral intentions for the audience. Thus, how these two parasocial phenomena affect audiences’ psychological and behavioral intentions in travel livestreaming is unknown. Previous research has explored various precedents for fostering * Corresponding author. E-mail addresses: [email protected] (Y. Lu), [email protected] (X. Liu), [email protected] (Y. Hu), [email protected] (C. Zhu). Contents lists available at ScienceDirect Journal of Hospitality and Tourism Management journal homepage: www.elsevier.com/locate/jhtm https://doi.org/10.1016/j.jhtm.2023.10.003 Received 3 June 2023; Received in revised form 29 September 2023; Accepted 1 October 2023

Journal of Hospitality and Tourism Management 57 (2023) 170–178 171 parasocial phenomena from an audience psychology perspective (Bi et al., 2021; Lim & Kim, 2011). However, from a social perspective, the characteristics and performance of media figures are also important factors influencing the occurrence of parasocial phenomena (Eyal & Rubin, 2003). According to interpersonal relationship theory, livestreamers and audiences can become closer through self-disclosure (Krasnova et al., 2010). The highly interactive form of livestreaming supports the formation of parasocial interactions and parasocial relationships by enhancing this sense of interpersonal intimacy (Deng et al., 2022). Therefore, the intimate self-disclosure of livestreamers may be a prerequisite for parasocial phenomena. Currently, few studies have examined how social media influencers’ intimate self-disclosure influences consumers’ purchase intentions. In the livestreaming environment, it must be seen whether influencers’ intimate self-disclosure has any effect. In addition, credibility is assessed in this study, which assesses how much audiences rely on and trust livestreamers’ information (Rogers & Bhowmik, 1970). Since trust is a deeper interpersonal relationship, Isaac and Grayson (2017) found that when audiences have higher credibility for livestreamers, the negative impact of livestreamers’ endorsements is reduced. Therefore, this study believes that in addition to the initial establishment of intimacy, attracting livestreamers and audiences to form a deeper trust relationship is of greater significance for the subsequent behavior and intention of audiences. To fill the research void above, this study proposed the following three questions: (1) Can parasocial interaction and parasocial relationships be effectively distinguished, and is there a causal relationship? (2) Do parasocial interaction and parasocial relationships mediate the relationship between intimate self-disclosure and purchase intention? (3) Will parasocial interaction, parasocial relationships, and credibility affect purchase intention? 2. Theoretical background and hypothesis development 2.1. Parasocial interaction theory Parasocial interaction theory is used to explain the imagined social relationships and interactions we have with media characters (Rasmussen, 2018). These social interactions and relationships resemble face-to-face social ties yet differ from them. Through this virtual interactive experience, the audience will form an attitude toward media figures and establish a stronger emotional bond with them (Hu et al., 2020). In a virtual environment, livestreaming provides a more favorable platform for live streamers and audiences to facilitate unprecedented interactions and connections. A relevant framework for examining the influence of livestreamers in travel livestreaming is provided by parasocial interaction theory (Chung & Cho, 2017). According to this theory, livestreamers have been shown to improve audience parasocial experience and stickiness through positive self-disclosure (Hu et al., 2020). Therefore, a travel livestreaming platform can employ livestreamers’ intimate self-disclosure to promote tourist marketing interaction, audience understanding, and trust in tourism products, thus enhancing audience intent to purchase tourism products. 2.2. Parasocial interaction and parasocial relationship Parasocial phenomena are used to summarize all different types of parasocial responses of audiences to media characters (Liebers & Schramm, 2019) and are one of the most popular and widespread research topics in the field of communication (Giles, 2002). Parasocial interaction and parasocial relationships, as the two most important phenomena of parasocial phenomena (McLaughlin & Wohn, 2021), were not distinguished when they were first proposed by Horton and Wohl (1956), which also led many scholars to follow up. In studies, the concepts and uses of these two concepts are often confused (Dibble et al., 2016; Liebers & Schramm, 2019; McLaughlin & Wohn, 2021), and they are even considered one phenomenon (Blight, 2016). Parasocial interaction refers to the illusory mutual awareness, attention, and adjustment between the audience and the media persona only during media exposure, specifically so that the audience feels that they are being directly addressed by the media persona (Dibble et al., 2016), which are immediate or short-term interactions (Sherrick et al., 2022). While parasocial relationships are short- or long-term (positive or negative) social relationships between the audience and the media persona (Deng et al., 2022; Giles, 2002), which may begin to develop during the period of media exposure or transcend media exposure (Dibble et al., 2016), the experience of parasocial interaction generally leads to the generation of parasocial relationships conceptually (Sherrick et al., 2022; Slater et al., 2018). Much of the current research on parasocial interaction is actually about parasocial relationships (Dibble et al., 2016), and researchers must be careful not to confuse the two concepts. 2.3. Intimate self-disclosure According to Kim and Song (2016), intimate self-disclosure is the intimate degree to which an individual discloses information about a particular area of his or her life to others. With the increasing convenience of sharing and communication on social media platforms, self-disclosure has become one of the most common behaviors in media communication (Jiang et al., 2010). In recent years, the research focus has increasingly shifted to how users view the self-disclosure of others, such as social media influencers (Chung & Cho, 2017). Bickart et al. (2015) found that the disclosure of intimate personal information has become a powerful tool for social media influencers to persuade consumers. When livestreaming the travel process and selling tourism products, livestreamers disclose intimate information to create a friendly and realistic environment. In this way, they compensate for the physical distance created by the virtual environment and reduce the psychological distance to the audience, thus influencing the audience’s feelings toward livestreamers and tourism products. As the livestreaming industry has led to the vigorous development of the livestreamer profession, the scope of social media influencers has gradually expanded. Although intimate self-disclosure was initially recognized as a marketing strategy commonly used by social media influencers, further research is needed in the livestreaming environment to verify the theoretical and practical impact of this strategy on tourism product marketing. 2.4. Hypothesis development 2.4.1. Intimate self-disclosure, parasocial interaction, and parasocial relationship Self-disclosure is an important aspect of all social interactions (Wang & Hu, 2022). Studies have shown that when YouTubers disclose personal information, the psychological distance between YouTubers and viewers will be shortened, and “face-to-face” interaction will be generated (Su et al., 2023). Therefore, livestreamers’ intimate self-disclosure will narrow the relationship with the audience and thus enhance parasocial interaction. Previous studies have confirmed that vloggers’ self-disclosure will lead viewers to a higher level of parasocial interaction (Kim & Song, 2016; Su et al., 2023; Wang & Hu, 2022). Therefore, we propose the following hypothesis: Hypothesis 1. Intimate self-disclosure is positively associated with parasocial interaction. Self-disclosure is an important prerequisite in all social relationships, and media characters or audiences often disclose information about themselves to engage and enhance interpersonal communication (Leite & Baptista, 2022; Munoz ˜ & Chen, 2023). The more self-disclosure there is in a relationship, the deeper the emotions involved in the relationship (Ferchaud et al., 2018), and these emotions often persist after media exposure. In the travel livestreaming process, the livestreamer’s Y. Lu et al.

Journal of Hospitality and Tourism Management 57 (2023) 170–178 172 self-disclosure leads to a closer relationship and liking between the two parties, and the parasocial relationship between the livestreamer and the audience increases accordingly. Previous studies have confirmed that social media influencers’ intimate self-disclosure plays a crucial role in the development of parasocial relationships with audiences (Leite & Baptista, 2022). Therefore, we propose the following hypothesis: Hypothesis 2. Intimate self-disclosure is positively associated with parasocial relationships. 2.4.2. Parasocial interaction and parasocial relationship Horton and Wohl (1956) argue that parasocial interaction is the moment when the audience feels as if they are in direct contact with the media characters, and the increase in interaction in such short moments will lead to long-term parasocial relations. This suggests that parasocial interaction and parasocial relationship are not only two different concepts but also potentially related (Sherrick, 2022). On social media platforms, interaction can create a sense of intimacy, connection, perceived friendship and understanding, and identification with celebrities (Chung & Cho, 2017). Labrecque’s (2014) research further proves that high interaction will enhance the parasocial relationship between consumers and brands. It can be inferred that under the background of travel livestreaming, the audience and the livestreamer further deepen their emotional connection through parasocial interaction, thus forming a deep parasocial relationship. Hence, we propose the following hypothesis: Hypothesis 3. Parasocial interaction is positively associated with parasocial relationships. 2.5. Intimate self-disclosure and credibility Credibility refers to the degree to which an information source is trustworthy and reliable (Rogers & Bhowmik, 1970) and is often used to reflect consumers’ trust in endorsers and the product information they provide (Baniya, 2017). Huang (2015) found that when bloggers disclose high-level intimate information, it increases readers’ familiarity with them, thereby forming cognitive and affective trust. Leite & Baptista, 2022 also confirmed through their research that social media influencers’ self-disclosure can enhance the audience’s positive evaluation and generate credibility. Hence, the following hypothesis is proposed: Hypothesis 4. Intimate self-disclosure is positively associated with credibility. 2.6. Parasocial interaction, parasocial relationship, credibility, and purchase intention Since the intimate self-disclosure of the livestreamer is unilateral, the formation of trust is based on interaction and relationships. First, tourism products usually have intangible characteristics. Due to the lack of visual observation of tourism products, the perceived uncertainty and risk of products are higher than those of tangible products (Lin et al., 2009). Therefore, travel livestreamers usually use more verbal descriptions and frequent interactions to reduce the audience’s uncertainty about products and enhance credibility (Reinikainen et al., 2020). Previous research has also demonstrated that parasocial interaction largely cultivates the loyalty and trust of influencers (Labrecque, 2014). According to parasocial interaction theory, when livestreamers engage in intimate self-disclosure, parasocial interaction ensues, which has a positive impact on enhancing audience trust. , two hypotheses are proposed: Hypothesis 5. Parasocial interaction is positively associated with credibility. Hypothesis 6. Parasocial interaction mediates the relationship between intimate self-disclosure and credibility. Second, in addition to interaction, interpersonal relationships are a key factor in enhancing credibility (Altman & Taylor, 1973). According to parasocial interaction theory, in the virtual interactive experience, the audience will form an attitude toward media figures and establish a stronger emotional bond with them (Hu et al., 2020). When audiences see the livestreamer as their “para-friend,” they have a higher degree of recognition and trust in the livestreamers (Perse & Rubin, 1989). Thus, when livestreamers deeply express themselves to the audience, the audience may develop a relationship with the livestreamers (e.g., para-friends) (Chung & Cho, 2017), resulting in a stronger sense of trust (Leite & Baptista, 2022). Therefore, two hypotheses are proposed: Hypothesis 7. Parasocial relationships are positively associated with credibility. Hypothesis 8. Parasocial relationships mediate the relationship between intimate self-disclosure and credibility. Purchase intention is defined as a consumer’s conscious plan or intent to buy a good or service, and it is also an indicator to measure the likelihood of purchasing consumption behavior (Fishbein & Ajzen, 1977). According to Lee and Watkins (2016), the brand cognition of consumers is favorably impacted by their parasocial reactions with vloggers. Many studies have proven that parasocial interaction is the premise of consumers’ purchase intention (Fazli-Salehi et al., 2022; Lee & Lee, 2022). Hence, we propose the following hypothesis: Hypothesis 9. Parasocial interaction is positively associated with purchase intention. Parasocial relationships have an important impact on online users of social media. Bi and Zhang’s (2023) research points out that because parasocial relationships resemble imagined friendships, consumers may make social comparisons with influencers. The motivation for this comparison may be self-improvement, which consumers can achieve by purchasing products (Lou & Kim, 2019). Previous scholars have confirmed that parasocial phenomena can affect consumers’ attitudes towards spokespersons and products (Gong & Li, 2017) and their purchase intentions for featured products (Sokolova & Kefi, 2020). Hence, we propose the following hypothesis: Hypothesis 10. Parasocial relationships are positively associated with purchase intention. According to credibility theory (Ohanian, 1990), customers are more likely to have good behavioral intentions when they believe the source more (Lau & Lee, 1999). If audiences perceive livestreamers as trustworthy, then they perceive the products recommended by livestreamers as having similar positive attributes (Spry et al., 2011). Higher credibility significantly influences purchasing intention, according to prior studies (Sokolova & Kefi, 2020), from which we conclude that credibility can increase audiences’ purchase intention. Hence, we propose the following hypothesis: Hypothesis 11. Credibility is positively associated with purchase intention. Based on the previous literature review and hypothesis development, we propose the following research model (Fig. 1). 3. Methodology 3.1. Study site and questionnaire design The main reason why this study chooses Ctrip Live (a well-known online travel agency in China) as the sample collection point is that it is currently one of the most popular travel broadcast outlets, with more than 1 million followers on TikTok alone. According to statistics, by the end of 2020, approximately 200 million consumers had booked travel Y. Lu et al.

Journal of Hospitality and Tourism Management 57 (2023) 170–178 173 while watching livestreaming from Ctrip, and the total volume of presale goods transactions had exceeded RMB 4 billion (Travel Daily, 2021). It is also noted that Ctrip Live has been followed by a large number of Chinese consumers in a short period of time, precisely because its co-founder James Liang’s personal IP has created an attraction. James often wear traditional Chinese folk clothes to match the theme of the destination during livestreams, which makes him close to consumers and makes him an online star. In 2020, he sold US$294 million in travel packages and hotel room bookings over 25 livestreams (Zhang, 2020). At present, Ctrip Live is also continuing to expand its influence through a stable team of official livestreamers. Therefore, Ctrip Live is a suitable sample collection point for this study. This study used empirically verified scales to avoid measurement bias from a single question. The measure of intimate self-disclosure is mainly based on Leite & Baptista, 2022. The credibility measure is from Sokolova and Kefi (2020). The measure of tourism product purchase intention was adapted from Alalwan (2018). In addition, since the difference between parasocial interaction and parasocial relationship is often confused in the literature, this adds to the confusion in the measurement of these two structures. However, Dibble et al. (2016) tested the scale of parasocial interaction by Rubin et al. (1985) and Hartmann & Goldhoorn, 2011 and found that the former effectively measured parasocial relationships, while the latter effectively measured parasocial interactions. Chen et al. (2022) and Reinikainen et al. (2020) used the two scales recommended by Dibble et al. (2016) to effectively study the parasocial interaction and parasocial relationship of online influencers. Therefore, in this study, the parasocial interaction measures are from Chen et al. (2022), and the parasocial relationship measures are from Reinikainen et al. (2020). All measures are adapted to the current research situation. A total of 32 items were measured using a 7-point Likert scale, of which 7 means completely agree and 1 means completely disagree. The research questionnaire consists of three parts. The first part is a screening question to determine the eligibility of participants by answering “Have you watched Ctrip livestreaming in the past two months? Only those who answered “yes” were allowed to complete the survey. The second part is used to measure the items in the five structures. The third part includes questions about the profile of the participants (gender, age, frequency of watching social media livestreams, income, education, and occupation). By first translating the English questionnaire into Chinese and then back into English, this study avoids translation bias. Then, the contents of the Chinese and English questionnaires were verified by inviting two travel professors. To further verify the validity of the questionnaire’s content before the end of January 2023, 50 samples were pretested. All the participants who participated in the pretest indicated that they could clearly understand the research questions of the questionnaire, so no additional revisions were made to the questionnaire. 3.2. Sample collection This study uses the online survey platform Tencent Questionnaire (https://wj.qq.com/) to distribute questionnaires and collect data. With more than 40 million users in China, it is one of the most widely used platforms for collecting survey data online. To reduce the possibility of variance in common methods, this study collected data at different times and through different channels (Tehseen et al., 2017). First, this study conducted purposive sampling, distributing online questionnaires to Ctrip’s live tik-tok fan base and official fan base at different times from February 1 to February 22, 2023, to ensure maximum access to valid samples. At the same time, to recruit more qualified samples, this study also conducted nonprobability convenience sampling, used the interest panel of the Tencent questionnaire to effectively find respondents who had watched Ctrip livestreaming, and sent questionnaire invitations by email. Previous studies have confirmed that online questionnaires using non-probability convenience sampling are reliable and representative (Zhu, Wu, Lu, Fong, & She, 2022). A total of 437 samples were collected in this study (pretest samples not included). 63 invalid questionnaires with consistent scores and incomplete answers were excluded, and 374 valid questionnaires were obtained. The validity was 85.58%. The effective samples were 51.1% female and 48.9% male. Most participants were 18–25 years old (57.8%), had a bachelor’s degree (51.9%), and earned RMB 5000 or less per month (51.6%). The student group made up 42% of the sample, which is expected since students have more free time and are one of the main social media users (Shewale, 2023). Similar to the China Live Streaming Industry Development Research Report, approximately 90% of participants watched social media livestreaming more than once or twice a week (iiMedia Research, 2022). The demographics of the respondents are basically consistent with the statistics of the Online Performance (Live and Short Video) Fig. 1. Proposed Theoretical Framework for this study. Y. Lu et al.

Journal of Hospitality and Tourism Management 57 (2023) 170–178 174 Industry Development Report for 2022–2023 (China Association of Performing Arts, 2023). 4. Results 4.1. Common method variance This study controls for common method variance (CMV) using these methods. First, the questionnaire outlined introduced the study’s purpose and assured the anonymity of the participants. Second, at the start of the survey, this study reassured respondents that there was no right or incorrect answer. Third, the response bias is reduced by randomly rotating the order of items in the scale. Finally, Harman’s single-factor test was employed to identify CMV following data collection. Principal component analysis showed that the first factor accounted for 46.38% of the total variance, less than the crucial limit of 50%, indicating that CMV was not significant in this study (Podsakoff et al., 2003). In addition, this study also checks the multicollinearity problem by the variance inflation factor (VIF). The results of VIF show that all VIF values are less than 3.3; thus, there is no collinearity (Kock & Lynn, 2012) (see Table 1). 4.2. The measurement model Table 2 shows the factor load and descriptive statistical analysis results of each measurable item, in which the PLS factor loading value of each item is greater than 0.7. Table 3 shows that Cronbach’s alpha and CR values of all variables are greater than 0.7, and AVE values are greater than 0.5, which shows that the structure has strong reliability and validity (Hair et al., 2010). In addition, the square root of the AVE of each construct is larger than its construct correlation, according to the results of the Fornell-Larcker criterion data (Fornell & Larcker, 1981), while the value of the heterotrait-monotrait ratio (HTMT) is less than 0.9 (Henseler et al., 2015), indicating the discriminant validity of the scale. 4.3. Hypothesis testing This study uses bootstrapping (5000 subsamples) to validate the hypothesized model. The results of the PLS-SEM are shown in Fig. 2. The results showed that livestreamers’ intimate self-disclosure significantly affected parasocial interaction (β = 0.489, p < 0.001) and parasocial relationships (β = 0.331, p < 0.001). Parasocial interaction significantly affected parasocial relationships (β = 0.589, p < 0.001). Both parasocial interaction and parasocial relationship were positively associated with credibility (β = 0.175, p < 0.001; β = 0.390, p < 0.001, respectively). Livestreamers’ intimate self-disclosure significantly affected credibility (β = 0.365, p < 0.001). Therefore, parasocial interaction and parasocial relationships may mediate the influences of livestreamers’ intimate selfdisclosure on credibility. Additionally, both parasocial relationships and credibility were positively associated with purchase intention (β = 0.478, p < 0.001; β = 0.268, p < 0.001, respectively), but parasocial interaction had no significant effect on purchase intention (β = 0.073, p = 0.191). Therefore, all hypotheses except H9 have been supported. R2 and Q2 were assessed in this study to assess the structural model. The findings revealed that the model has strong explanatory power, with R2 values of parasocial interaction, parasocial relationship, credibility, and purchase intention of 0.240, 0.646, 0.657, and 0.575, respectively, and Q2 values of 0.152, 0.425, 0.377, and 0.422, respectively. This demonstrates the model’s adequate predictive relevance (Hair, Matthews, Matthews, & Sarstedt, 2017). PLS bootstrapping (5000 subsamples) was used in this paper to confirm the mediating role of parasocial interaction and parasocial relationships. The results showed that parasocial interaction (indirect effect = 0.085, CI = [0.039, 0.136]) and parasocial relationship (indirect effect = 0.129, CI = [0.084, 0.175]) played a significant mediating role between livestreamers’ intimate self-disclosure and credibility (as shown in Table 4); therefore, H6 and H8 were supported. According to the mediation classification developed by Zhao et al. (2010), parasocial interaction and parasocial relationship were complementary mediations in livestreamers’ intimate self-disclosure and credibility, since both direct effects and mediating effects existed in the same direction. 5. Discussion 5.1. The impact of livestreamers’ intimate self-disclosure on parasocial interaction and parasocial relationships Research has found that livestreamers’ intimate self-disclosure leads to higher levels of parasocial interaction and parasocial relationships in the audience. On the one hand, when livestreamers engage in intimate self-disclosure, the audience will have a favorable impression of the sincerity of the livestreamers, and parasocial interaction will be generated during the livestreaming process, which is consistent with the research results of Wang & Hu, 2022. On the other hand, since self-disclosure often expresses liking, livestreamers’ intimate self-disclosure will make the relationship between the two parties closer, and the parasocial relationship will also be enhanced, which also validates the statement of Leite & Baptista, 2022. In addition, this study also found that parasocial interaction significantly affected parasocial relationships, which is consistent with the findings of Sherrick et al. (2022). Specifically, as the parasocial interaction between livestreamers and the audience increases, the parasocial relationship generated by both parties gradually increases. Previously, most scholars only confirmed the relationship between self-disclosure and parasocial interaction and rarely paid attention to the parasocial relationship. Based on the background of travel livestreaming, this study explains that livestreamers’ intimate selfdisclosure generates two different types of parasocial phenomena, which fills the research gap. 5.2. Mediating role of parasocial interaction and parasocial relationship The results showed that parasocial interaction played a mediating role between intimate self-disclosure and credibility. Since the perceived Table 1 Sample profile (n = 374). Frequency Percent Gender Males 183 48.9 Females 191 51.1 Age 18–25 216 57.8 26–30 76 20.3 31–40 63 16.8 41–50 16 4.3 51–60 3 0.8 Education Background High school or below 29 7.8 Technical secondary school or junior college 99 26.5 Undergraduate degree 194 51.9 Master’s degree and above 52 13.9 Personal Monthly Income (RMB) ≤5000 193 51.6 5001–15000 148 39.6 15001-35000 29 7.8 >35000 4 1.1 Occupation Students 157 42 National public officials 17 4.5 Public institutions personnel 52 13.9 Private enterprise employees 74 19.8 Businessmen 6 1.6 Farmers 4 1.1 Freelancers 48 12.8 Others 16 4.3 Frequency of watching live streams Almost every day 181 48.4 Once or twice a week 151 40.4 Once or twice a month 36 9.6 Rarely watch 6 1.6 Y. Lu et al.

Journal of Hospitality and Tourism Management 57 (2023) 170–178 175 Table 2 Descriptive statistics and factor loadings. Construct Abbreviation Items Mean Standard deviation Factor loading Intimate selfdisclosure ISD1 The livestreamer shares their information about themselves in the travel livestreaming. 5.767 1.0391 0.755 ISD2 The livestreamer talks about their behaviors in the travel livestreaming. 5.372 1.0779 0.754 ISD3 The livestreamer shares their feelings in the travel livestreaming. 5.316 1.1282 0.760 ISD4 The livestreamer shares their emotions in the travel livestreaming. 5.647 1.0553 0.735 ISD5 The livestreamer shares their desires in the travel livestreaming. 5.690 1.0459 0.798 ISD6 The livestreamer talks about their moods in the travel livestreaming. 5.433 1.0908 0.740 ISD7 The livestreamer shares their thoughts in the travel livestreaming. 5.717 0.9905 0.761 ISD8 The livestreamer shares their opinions in the travel livestreaming. 5.249 1.0685 0.746 ISD9 The livestreamer shares their beliefs in the travel livestreaming. 5.265 1.0846 0.718 Parasocial Interaction PSI1 While watching the travel livestreaming,I had the feeling that the livestreamer was aware of me. 5.390 1.0265 0.774 PSI2 While watching the travel livestreaming,the livestreamer knew I was there. 5.299 1.0718 0.775 PSI3 While watching the travel livestreaming,the livestreamer knew I was aware of him/her. 5.176 1.1061 0.861 PSI4 While watching the travel livestreaming,the livestreamer knew I paid attention to him/her. 5.056 1.2119 0.784 PSI5 While watching the travel livestreaming,the livestreamer knew that I reacted to him/her. 5.439 1.0434 0.823 PSI6 While watching the travel livestreaming,the livestreamer reacted to what I said or did. 4.430 1.3929 0.804 Parasocial Relationship PSR1 I look forward to watching the livestreamer again. 4.227 1.5392 0.833 PSR2 If the livestreamer in live Streaming appeared on another channel, I would watch that livestreaming. 4.687 1.3625 0.785 PSR3 When I am watching the livestreamer, I feel as if I am part of his/her group. 5.112 1.2332 0.829 PSR4 I think the live streamer is like an old friend. 5.104 1.3088 0.845 PSR5 I would like to meet the livestreamer in person. 5.027 1.2315 0.797 PSR6 If there was a story about the livestreamer in a newspaper or magazine, I would read it. 5.201 1.1056 0.849 PSR7 The livestreamer makes me feel comfortable, as if I am with friends. 5.110 1.1897 0.831 PSR8 When the livestreamer shows me how he/she feels about the travel product, it helps me make up my own judgment about the product. 4.992 1.2021 0.744 Credibility C1 I find this livestreamer expert in his/her domain. 4.741 1.3138 0.814 C2 I find this livestreamer efficient in his/her job. 4.489 1.5144 0.837 C3 I find this livestreamer trustworthy. 4.802 1.3010 0.869 C4 I think this livestreamer cares about his/her followers. 4.941 1.1793 0.818 C5 This livestreamer updates regularly her livestream content. 5.393 1.0525 0.733 Purchase Intention PI1 I will buy the travel products recommended by that livestreamer. 4.701 1.2980 0.871 PI2 I desire to buy the travel products recommended by that livestreamer. 4.770 1.2190 0.846 PI3 I am likely to buy the travel products recommended by that livestreamer. 4.928 1.2652 0.868 PI4 I plan to purchase the travel products recommended by that livestreamer. 4.743 1.3436 0.866 Table 3 Reliability, construct validity, and correlation. Cronbach’s Alpha CR AVE Fornell-Larcker Criterion Heterotrait-Monotrait Ratio C ISD PSI PSR PI C ISD PSI PSR PI Credibility(C) 0.873 0.908 0.665 0.815 Intimate self-disclosure (ISD) 0.904 0.921 0.566 0.692 0.752 0.778 Parasocial Interaction (PSI) 0.891 0.916 0.647 0.646 0.489 0.804 0.726 0.533 Parasocial Relationship (PSR) 0.927 0.940 0.664 0.747 0.619 0.751 0.815 0.831 0.670 0.820 Purchase Intention (PI) 0.886 0.921 0.744 0.673 0.489 0.605 0.733 0.863 0.761 0.536 0.679 0.807 Fig. 2. Results of PLS-SEM analysis. Y. Lu et al.

Journal of Hospitality and Tourism Management 57 (2023) 170–178 176 risk and uncertainty of purchasing travel products online is higher than that of purchasing tangible products (Lin et al., 2009), the parasocial interaction between livestreamers and the audience often plays an important role in enhancing credibility (Reinikainen et al., 2020). Therefore, when a travel livestreamer discloses more private information, parasocial interaction will increase, and credibility will improve. The study also found that parasocial relationships also play a mediating role between intimate self-disclosure and credibility. When travel livestreamers’ intimate self-disclosure establishes a deep interpersonal relationship (e.g., parasocial relationship) with the audience, the audience will have a stronger sense of trust in the livestreamer. This finding supports the idea of Leite & Baptista, 2022. This study confirms that parasocial interaction and parasocial relationships play an important role in the relationship between intimate self-disclosure and credibility in travel livestreaming, providing empirical evidence for follow-up research. 5.3. The effects of parasocial relationships and credibility on purchase intention According to the study, credibility and parasocial relationships significantly affected purchase intention. This finding is in line with those of Bi and Zhang (2023) and Sokolova and Kefi (2020). With the promotion of the parasocial relationship, the audience’s recognition of the livestreamer will continue to increase, thus promoting the audience’s purchase intention. Meanwhile, the improvement of credibility will also enhance the audience’s purchase intention. In addition, contrary to the findings of Lee and Lee (2022), Sokolova and Kefi (2020), Fazli-Salehi et al. (2022), and Shen et al. (2022), this study showed no evidence of a significant relationship between parasocial interaction and purchase intention. This may be due to the following reasons. First, previous scholars have mainly studied the relationship between Instagram, YouTube, Facebook, etc., influencers and audience parasocial interaction and purchase intention, but livestreaming allows audio and video integration with online platforms and real-time broadcasting through streaming livestreamers and audiences to communicate simultaneously. Although livestreaming is more conducive to parasocial interaction, it merely improves audience information intake and does not affect purchase intention. Only trust in the information or livestreamer will affect the audience’s attitude (Sokolova & Kefi, 2020). Finally, this study is based on the background of travel livestreaming (mainly selling group tours, customized tours, etc.). Because tourism products not only have intangible features but also usually involve complex choices and higher costs, the perceived risk and uncertainty of online shopping for tourism products is often high (Lin et al., 2009). Therefore, in travel livestreaming, parasocial interaction often fails to promote purchase intentions, which may explain the discrepancy with Shen et al. (2022). 6. Conclusion and limitations 6.1. Theoretical contribution First, in the context of travel livestreaming, this study enriches the discussion on parasocial relationships and parasocial interactions. Currently, research on parasocial phenomena in travel livestreaming is relatively limited, with only a few studies exploring parasocial interaction (Shen et al., 2022), and no scholars have investigated parasocial interaction and parasocial relationships simultaneously. Our study fills this gap. The current study found that when a travelling livestreamer engages in intimate self-disclosure, it can help audiences form two types of parasocial phenomena (parasocial interaction and parasocial relationship), which provides an important theoretical basis for revealing the relationship between media influencers’ intimate self-disclosure and parasocial phenomena and enriches the field’s interpretation of parasocial interaction theory. Second, this study helps researchers understand the further application of parasocial interaction theory in travel livestreaming from the perspectives of parasocial interaction and parasocial relationships. Although early studies contributed to the development of parasocial interaction and parasocial relationships (e.g., Leite & Baptista, 2022; Shen et al., 2022), these studies mainly focused on a particular type of parasocial phenomenon (e.g., parasocial interaction) to explain how purchase intention arises. Previous studies have theoretically explained the differences in the concepts and usages of parasocial interaction and parasocial relationship (Deng et al., 2022; Dibble et al., 2016), and this study provides further evidence for the differences in concept and use between parasocial interaction and parasocial relationship. In addition, this study fills the gap left by Deng et al. (2022) based on travel livestreaming. Although Deng et al. (2022) used a qualitative approach to explain the antecedents of the parasocial phenomenon, the consequences of parasocial interaction and parasocial relationship formation were not examined, and our work thus addresses this research gap. Finally, the result that parasocial interaction has no significant effect on purchase intention contradicts previous research results (Fazli-Salehi et al., 2022; Lee & Lee, 2022), which has important implications for tourism marketing research. This finding supports credibility theory (Ohanian, 1990), and only when the audience has more trust in the source is it possible to induce them to have some positive behavioral intentions. Studies have shown that the interaction between the travel livestreamer and the audience can only increase the audience’s intake of relevant information about tourism products but fails to generate trust in the audience, which may be the reason why they have no purchase intention. Since parasocial interaction may not always be the primary factor enhancing purchase intention, marketing research should take this into account. Future research could examine the relationship between parasocial interaction and purchase intention in other livestreaming contexts or environments. 6.2. Practical contributions The study’s findings have applications for the marketing of tourism, livestreaming platforms, and tourism livestreaming agencies. First, it should be considered that the livestreamer’s intimate self-disclosure behavior promotes parasocial interaction and parasocial relationships with the audience. The findings of this study show that intimate selfdisclosure should be a key component of a livestreaming marketing plan to engage the audience and establish a sense of reality. Therefore, practitioners should consider how to allow livestreamers to better reveal themselves on the livestreaming platform, not just during product introductions. For example, intimate self-disclosure can be transformed into a livestreamer’s marketing technique through professional training that integrates product information with the personal information disclosed by livestreamers as a source of advice to the audience. Second, the results of this study also showed a causal relationship between parasocial interaction and parasocial relationship. It is suggested that travel livestreaming practitioners should focus on real-time interaction and relationship building with audiences. For example, in travel livestreaming, it can be considered to encourage the active participation of the audience by improving the reward mechanism Table 4 The mediation effect of parasocial interaction and parasocial relationship. Hypothesized relationships Direct effects Indirect effects Mediation livestreamers’ intimate selfdisclosure→parasocial interaction→credibility 0.365 (0.270, 0.462) 0.086 (0.039, 0.136) Complementary mediation livestreamers’ intimate selfdisclosure→parasocial relationship→credibility 0.365 (0.270, 0.462) 0.129 (0.084, 0.175) Complementary mediation Note: Values in parentheses are 95% confidence intervals (CI). Y. Lu et al.

Journal of Hospitality and Tourism Management 57 (2023) 170–178 177 during livestreaming, and livestreamers should always pay attention to strengthening the real-time response to the audience. The audience can be rated by monitoring the viewing duration or number of audiences. For audiences with higher ratings, livestreamers should give special treatment and respond to establish a good parasocial relationship. This creates a “frequent visitor” effect. Finally, the findings demonstrate that parasocial relationships and credibility significantly influence purchase intention. In travel livestreaming, it is very important for the livestreamer and the audience to establish deep interpersonal relationships and trust. Relevant practitioners should pay full attention to the content and attitude of the audience. The livestreaming team and livestreamers should quickly recognize the audience’s doubts about the product or organization and respond and resolve them properly to eliminate the audience’s uncertainty about the information and strengthen the audience’s trust in the livestreaming content. Only by further establishing a higher degree of interpersonal relationship with the audience can the audience’s purchase intention be effectively improved. 6.3. Limitations and future research There are several research deficiencies in the current study. First, this study mainly analysed the mechanism of the impact of livestreamers’ intimate self-disclosure on purchase intention and did not consider the influence of the celebrity effect. Therefore, future research could consider grouping celebrity and noncelebrity livestreamers for comparison. Second, this study does not consider distinguishing whether the purchase intention generated by the audience is “future” or immediate.” Because tourism products are different from other commodities, they need to consider factors such as time, cost, and transportation. Therefore, future research can build on this to analyse the responses generated by the audience. Third, since parasocial relationships are a dynamic development process (Deng et al., 2022) and this study is a cross-sectional study, future research can examine the development and dynamics of parasocial relationships over a long period of time. Fourth, the tourism products sold by livestreamers include group tours, customized tours, hotels, scenic spot tickets, etc., but this study did not subdivide them. Due to the different tourism products of livestreaming, there may be differences in the research results (Shen et al., 2022). Therefore, future research can refine it. References Alalwan. (2018). Investigating the impact of social media advertising features on customer purchase intention. International Journal of Information Management, 42, 65–77. https://doi.org/10.1016/j.ijinfomgt.2018.06.001 Altman, I., & Taylor, D. A. (1973). Social penetration: The development of interpersonal relationships. Holt, Rinehart, & Winston. Baniya, R. (2017). Components of celebrity endorsement affecting brand loyalty of Nepali customers. Journal of Business Management and Research, 2(1–2), 52–65. https ://doi-org.libezproxy.must.edu.mo/10.3126/jbmr.v2i1-2.18151. Bickart, B., Kim, S., Pai, S., & Brunel, F. (2015). How social media influencers build a brand following by sharing secrets. In S. Fournier, M. Breazeale, & J. Avery (Eds.), Strong brands, strong relationships (pp. 172–184). New York: Routledge. Bi, Y., Yin, J., & Kim, I. (2021). Fostering a young audience’s media-induced travel intentions: The role of parasocial interactions. Journal of Hospitality and Tourism Management, 47, 398–407. https://doi.org/10.1016/j.jhtm.2021.04.011 Bi, & Zhang, R. (2023). “I will buy what my ‘friend’ recommends”: The effects of parasocial relationships, influencer credibility and self-esteem on purchase intentions. The Journal of Research in Indian Medicine, 17(2), 157–175. https://doi. org/10.1108/JRIM-08-2021-0214 Blight. (2016). Relationships to video game streamers: Examining gratifications, parasocial relationships, fandom, and community affiliation online. ProQuest Dissertations Publishing. Chen, X., Hyun, S. S., & Lee, T. J. (2022). The effects of parasocial interaction, authenticity, and self-congruity on the formation of consumer trust in online travel agencies. International Journal of Tourism Research, 24(4), 563–576. https://doi.org/ 10.1002/jtr.2522 China Association of Performing Arts. (2023). Online performance (live and short video) industry development Report. https://dzswgf.mofcom.gov.cn/news/43/2023/5 /1685338992909.html. (Accessed 18 September 2023). Chung, S., & Cho, H. (2017). Fostering parasocial relationships with celebrities on social media: Implications for celebrity endorsement. Psychology and Marketing, 34(4), 481–495. https://doi.org/10.1002/mar.21001 Deng, Z., Benckendorff, P., & Wang, J. (2022). From interaction to relationship: Rethinking parasocial phenomena in travel live streaming. Tourism Management, 93, Article 104583. https://doi.org/10.1016/j.tourman.2022.104583 Dibble, J. D., Hartmann, T., & Rosaen, S. F. (2016). Parasocial interaction and parasocial relationship: Conceptual clarification and a critical assessment of measures. Human Communication Research, 42(1), 21–44. https://doi.org/10.1111/hcre.12063 Eyal, & Rubin, A. M. (2003). Viewer aggression and homophily, identification, and parasocial relationships with television characters. Journal of Broadcasting & Electronic Media, 47(1), 77–98. https://doi.org/10.1207/s15506878jobem4701_5 Fazli-Salehi, R., Jahangard, M., Torres, I. M., Madadi, R., & Zúniga, ˜ M.A. ´ (2022). Social media reviewing channels: The role of channel interactivity and vloggers’ selfdisclosure in consumers’ parasocial interaction. Journal of Consumer Marketing, 39 (2), 242–253. https://doi.org/10.1108/JCM-06-2020-3866 Ferchaud, A., Grzeslo, J., Orme, S., & LaGroue, J. (2018). Parasocial attributes and YouTube personalities: Exploring content trends across the most subscribed YouTube channels. Computers in Human Behavior, 80, 88–96. https://doi.org/10.1016/j. chb.2017.10.041 Fishbein, M., & Ajzen, I. (1977). Belief, attitude, intention, and behavior: An introduction to theory and research. Philosophy and Rhetoric, 10(2), 177–188. https://doi.org/ 10.1177/002224378101800104 Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18, 39–50. https://doi.org/10.1177/002224378101800104 Giles, D. C. (2002). Parasocial interaction: A review of the literature and a model for future research. Media Psychology, 4, 279–305. https://doi.org/10.1207/S1532785. XMEP0403_04 Gong, W., & Li, X. (2017). Engaging fans on microblog: The synthetic influence of parasocial interaction and source characteristics on celebrity endorsement. Psychology and Marketing, 34(7), 720–732. https://doi.org/10.1002/mar.21018 Hair, J. F., Black, W. C., Babin, B. J., & Anderson, R. E. (2010). Multivariate data analysis: A global perspective. Hoboken, NJ: Prentice Hall. Hair, J. F., Jr., Matthews, L. M., Matthews, R. L., & Sarstedt, M. (2017). PLS-SEM or CBSEM: Updated guidelines on which method to use. International Journal of Multivariate Data Analysis, 1(2), 107–123. https://doi.org/10.1504/ IJMDA.2017.087624 Hartmann, T., & Goldhoorn, C. (2011). Horton and Wohl revisited: Exploring viewers’ experience of parasocial interaction. Journal of Communication, 61(6), 1104–1121. https://doi.org/10.1111/j.1460-2466.2011.01595.x Henseler, J., Ringle, C. M., & Sarstedt, M. (2015). A new criterion for assessing discriminant validity in variance-based structural equation modeling. Journal of the Academy of Marketing Science, 43, 115–135. https://doi.org/10.1007/s11747-014- 0403-8 Horton, D., & Wohl, R. (1956). Mass communication and para-social interaction: Observations on intimacy at a distance. Psychiatry, 19(3), 215–229. Huang, L. (2015). Trust in product review blogs: The influence of self-disclosure and popularity. Behaviour & Information Technology, 34(1), 33–44. doi: 10.1080 /0144929X.2014.978378. Hu, L., Min, Q., Han, S., & Liu, Z. (2020). Understanding followers’ stickiness to digital influencers: The effect of psychological responses. International Journal of Information Management, 54. doi: 10.1016/j.ijinfomgt.2020.102169. Isaac, M. S., & Grayson, K. (2017). Beyond skepticism: Can accessing persuasion knowledge bolster credibility? Journal of Consumer Research, 43(6), 895–912. https://doi.org/10.1093/jcr/ucw063 Jiang, L., Bazarova, N., & Hancock, J. (2010). The disclosure-intimacy link in computer mediated communication: An attributional extension of the hyperpersonal model. Human Communication Research, 37(1), 58–77. https://doi.org/10.1111/j.1468- 2958.2010.01393.x Kim, J., & Song, H. (2016). Celebrity’s self-disclosure on twitter and parasocial relationships: A mediating role of social presence. Computers in Human Behavior, 62, 570–577. https://doi.org/10.1016/j.chb.2016.03.083 Kock, N., & Lynn, G. (2012). Lateral collinearity and misleading results in variance-based SEM: An illustration and recommendations. Journal of the Association for information Systems, 13(7). https://doi.org/10.17705/1jais.00302 Krasnova, H., Spiekermann, S., Koroleva, K., & Hildebrand, T. (2010). Online social networks: Why we disclose. Journal of Information Technology, 25(2), 109–125. https://doi.org/10.1057/jit.2010.6 Labrecque, L. (2014). Fostering consumer–brand relationships in social media environments: The role of parasocial interaction. Journal of Interactive Marketing, 28 (2), 134–148. https://doi.org/10.1016/j.intmar.2013.12.003 Lau, G. T., & Lee, S. H. (1999). Consumers’ trust in a brand and the link to brand loyalty. Journal of Market-Focused Management, 4(4), 341–370. https://doi.org/10.1023/A: 1009886520142 Lee, M., & Lee, H. (2022). Do parasocial interactions and vicarious experiences in the beauty YouTube channels promote consumer purchase intention? International Journal of Consumer Studies, 46(1), 235–248. https://doi.org/10.1111/ijcs.12667 Lee, J. E., & Watkins, B. (2016). YouTube vloggers’ influence on consumer luxury brand perceptions and intentions. Journal of Business Research, 69, 5753–5760. https://doi. org/10.1016/j.jbusres.2016.04.171 Leite, F. P., & Baptista, P. D. P. (2022). The effects of social media influencers’ selfdisclosure on behavioral intentions: The role of source credibility, parasocial relationships, and brand trust. Journal of Marketing Theory and Practice, 30(3), 295–311. https://doi.org/10.1080/10696679.2021.1935275 Y. Lu et al.

Journal of Hospitality and Tourism Management 57 (2023) 170–178 178 Liebers, N., & Schramm, H. (2019). Parasocial interactions and relationships with media Characters–An inventory of 60 years of research. Communication Research Trends, 38 (2), 4–31. Lim, C. M., & Kim, Y. K. (2011). Older consumers’ Tv home shopping: Loneliness, parasocial interaction, and perceived convenience. Psychology and Marketing, 28(8), 763–780. https://doi.org/10.1002/mar.20411 Lin, P. J., Jones, E., & Westwood, S. (2009). Perceived risk and risk-relievers in online travel purchase intentions. Journal of Hospitality Marketing & Management, 18(8), 782–810. doi: 10.1080/19368620903235803. Lou, C., & Kim, H. K. (2019). Fancying the new rich and famous? Explicating the roles of influencer content, credibility, and parental mediation in adolescents’ parasocial relationship, materialism, and purchase intentions. Frontiers in Psychology, 10, 2567. https://doi.org/10.3389/fpsyg.2019.02567 Lueck, J. A. (2015). Friend-zone with benefits: The parasocial advertising of Kim Kardashian. Journal of Marketing Communications, 21(2), 91–109. https://doi.org/ 10.1080/13527266.2012.726235 McLaughlin, C., & Wohn, D. Y. (2021). Predictors of parasocial interaction and relationships in live streaming. Convergence, 27(6), 1714–1734. https://doi.org/ 10.1177/13548565211027807 Munoz, ˜ K. E., & Chen, L.-H. (2023). Can dating app users’ self-disclosure foster travel intentions? An appnography approach. Journal of Hospitality and Tourism Management, 55, 493–501. https://doi.org/10.1016/j.jhtm.2023.05.020 Ohanian, R. (1990). Construction and validation of a scale to measure celebrity endorsers’ perceived expertise, trustworthiness, and attractiveness. Journal of Advertising, 19(3), 39–52. https://doi.org/10.1080/00913367.1990.10673191 Perse, E. M., & Rubin, R. B. (1989). Attribution in social and parasocial relationships. Communication Research, 16(1), 59–77. https://doi.org/10.1177/ 009365089016001003 Podsakoff, P. M., MacKenzie, S. B., Lee, J.-Y., & Podsakoff, N. P. (2003). Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88(5), 879–903. https://doi.org/10.1037/ 0021-9010.88.5.879 Rasmussen, L. (2018). Parasocial interaction in the digital age: An examination of relationship building and the effectiveness of YouTube celebrities. The Journal of Social Media in Society Spring, 7(1), 280–294. Reinikainen, H., Munnukka, J., Maity, D., & Luoma-aho, V. (2020). “You really are a great big sister” - parasocial relationships, credibility, and the moderating role of audience comments in influencer marketing. Journal of Marketing Management, 36 (3–4), 279–298. https://doi.org/10.1080/0267257X.2019.1708781 iiMedia Research. (2022). 2021 China online live streaming industry development research Report. https://www.iimedia.cn/c400/83735.html (accessed March 24, 2023). Rogers, E. M., & Bhowmik, D. K. (1970). Homophily-heterophily: Relational concepts for communication research. Public Opinion Quarterly, 34(4), 523–538. https://doi.org/ 10.1086/267838 Rubin, A. M., Perse, E. M., & Powell, R. A. (1985). Loneliness, parasocial interaction, and local television news viewing. Human Communication Research, 12(2), 155–180. Shen, H., Zhao, C., Fan, D. X. F., & Buhalis, D. (2022). The effect of hotel livestreaming on viewers’ purchase intention: Exploring the role of parasocial interaction and emotional engagement. International Journal of Hospitality Management, 107, Article 103348. https://doi.org/10.1016/j.ijhm.2022.103348 Sherrick, B., Smith, C., Jia, Y., Thomas, B., & Franklin, S. B. (2022). How parasocial phenomena contribute to sense of community on twitch. Journal of Broadcasting & Electronic Media, 67(1), 47–67. https://doi.org/10.1080/08838151.2022.2151599 Shewale, R. (2023). Social media users global demographics. https://www.demandsage. com/social-media-users/. (Accessed 18 September 2023). Slater, M. D., Ewoldsen, D. R., & Woods, K. W. (2018). Extending conceptualization and measurement of narrative engagement after-the-fact: Parasocial relationship and retrospective imaginative involvement. Media Psychology, 21(3), 329–351. doi: 10.1 080/15213269.2017.1328313. Sokolova, K., & Kefi, H. (2020). Instagram and YouTube bloggers promote it, why should I buy? How credibility and parasocial interaction influence purchase intentions. Journal of Retailing and Consumer Services, 53, Article 101742. https://doi.org/ 10.1016/j.jretconser.2019.01.011 Spry, A., Pappu, R., & Bettina Cornwell, T. (2011). Celebrity endorsement, brand credibility and brand equity. European Journal of Marketing, 45(6), 882–909. https:// doi.org/10.1108/03090561111119958 Su, B. C., Wu, L. W., & Wu, J. P. (2023). Exploring the characteristics of YouTubers and their influence on viewers’ purchase intention: A viewers’ pseudo-social interaction perspective. Sustainability, 15(1), 550. https://doi.org/10.3390/su15010550 Tehseen, S., Ramayah, T., & Sajilan, S. (2017). Testing and controlling for common method variance: A review of available methods. Journal of management sciences, 4 (2), 142–168. https://doi.org/10.20547/jms.2014.1704202 Travel Daily. (2021). Ctrip live: 200 million people planted "unknown travel", and more than 3,000 hotels and scenic spots joined. https://www.traveldaily.cn/articl e/142685. (Accessed 24 March 2023). Tsay-Vogel, M., & Schwartz, M. L. (2014). Theorizing parasocial interactions based on authenticity: The development of a media figure classification scheme. Psychology of Popular Media Culture, 3(2), 66–78. https://doi.org/10.1037/a0034615 Wang, E. S., & Hu, F. T. (2022). Influence of self-disclosure of internet celebrities on normative commitment: The mediating role of para-social interaction. The Journal of Research in Indian Medicine, 16(2), 292–309. https://doi.org/10.1108/JRIM-09- 2020-0194 Xie, C., Yu, J., Huang, S. S., & Zhang, J. (2022). Tourism e-commerce live streaming: Identifying and testing a value-based marketing framework from the live streamer perspective. Tourism Management, 91, Article 104513. https://doi.org/10.1016/j. tourman.2022.104513 Yang, J., Zeng, Y., Liu, X., & Li, Z. (2022). Nudging interactive cocreation behaviors in live-streaming travel commerce: The visualization of real-time danmaku. Journal of Hospitality and Tourism Management, 52, 184–197. https://doi.org/10.1016/j. jhtm.2022.06.015 Yang, J., Zhang, D., Liu, X., Hua, C., & Li, Z. (2022). Destination endorsers raising on short-form travel videos: Self-image construction and endorsement effect measurement. Journal of Hospitality and Tourism Management, 52, 101–112. https:// doi.org/10.1016/j.jhtm.2022.06.003 Zhang, J. (2020). Live-streaming here to stay as Chinese travel business almost ‘fully recovers’ from Covid-19, Trip. com co-founder says. South China Morning Post htt ps://www.scmp.com/tech/tech-leaders-and-founders/article/3104532/live-stre aming-here-stay-chinese-travel-business. (Accessed 11 April 2023). Zhao, X., Lynch, J. G., Jr., & Chen, Q. (2010). Reconsidering baron and kenny: Myths and truths about mediation analysis. Journal of Consumer Research, 37(2), 197–206. https://doi.org/10.1086/651257 Zheng, S., Wu, M., & Liao, J. (2023). The impact of destination live streaming on viewers’ travel intention. Current Issues in Tourism, 26(2), 184–198. https://doi.org/10.1080/ 13683500.2022.2117594 Zhu, C., Wu, D. C. W., Lu, Y., Fong, L. H. N., & She, L. S. (2022). When Virtual Reality meets destination marketing: The mediating role of presences between vividness and user responses. Journal of Vacation Marketing. https://doi.org/10.1177/ 13567667221141414. ISSN: 1356-7667 eISSN: 1479-1870. Y. Lu et al.

Journal of Hospitality and Tourism Management 57 (2023) 250–257 Available online 27 October 2023 1447-6770/© 2023 The Authors. Published by Elsevier Ltd. on behalf of CAUTHE - COUNCIL FOR AUSTRALASIAN TOURISM AND HOSPITALITY EDUCATION. All rights reserved. Job searching during the pandemic: The roles of job search constraints, stress, and coping on industry turnover intentions Iuliana Popa, Juan M. Madera * The Conrad N. Hilton College of Global Hospitality Leadership, The University of Houston, United States 1. Introduction In March of 2020, the US economy experienced a significant downturn with the spread of the COVID-19 pandemic. Due to pandemic concerns and lockdowns, many jobs were lost across various sectors; however, the hospitality industry was by far the hardest-hit industry in the US, with a loss of over 8.2 million jobs between February and April of 2020 (U.S. Bureau of Labor Statistics, 2021). Two-thirds of all restaurant employees in the US had lost their jobs by April 2020 (National Restaurant Association, 2020), while around the same time, eight out of every ten hotel rooms were empty, causing 2020 to be forecasted as the worst year for occupancy ever recorded (National Restaurant Association, 2020). Despite vaccine rollouts since, the hospitality industry is still facing a labor shortage in the midst of both increasing wages and turnover (Liu & DeMicco, 2021; Maduro, 2022). One reason for this labor shortage is that the pandemic has led hospitality talent to leave the industry given its response and vulnerability to the pandemic (King et al., 2021). The hospitality industry is labor intensive, relying on frontline customer service employees. However, the COVID-19 pandemic led to a reduction of labor via layoffs, furloughs, and closing of businesses (Yu, Lee, & Madera, 2021). Those who remained employed also experienced a lack of guidance from employers, changes in the operation of business and work conditions that have resulted in stress among employees (Guzzo et al., 2021; Han et al., 2021). For instance, stress related to working during the COVID-19 pandemic can also produce negative employee attitudes towards the hospitality industry as an employer, such as intentions to leave the hospitality industry (e.g., Chen & Chen, 2021; Yu, Lee, & Madera, 2021). Lastly, the pandemic has also been linked to job insecurity among employed talent (Abbas et al., 2021). For example, Chen and Eyoun (2021) found that fear of the COVID-19 virus related to feelings of job insecurity among restaurant employees. Although much has been learned about how the industry’s response to the pandemic (i.e., labor reduction and changes to the work environment) has led to employees’ intentions to leave the industry (Chen & Chen, 2021; King et al., 2021; Yu, Lee, & Madera, 2021), less known is how the pandemic has affected those in the hospitality job market. Specifically, there is a dearth of knowledge related to how the pandemic has also constrained the job search process and subsequently hospitality job candidate attitudes about working in the industry. Job search constraints has been defined as “situational factors that might limit or restrict an individual’s job search efforts” (Wanberg et al., 1999, p. 899). Constraints might include personal/family responsibilities, financial difficulties, and difficulties with the job application process, which can negatively impact job seekers’ efforts during the job search process. Despite these realities, very little is known about how COVID-19 related job search constraints might motivate hospitality talent to leave the industry. This is an unfortunate but important gap in the literature, given the labor shortage the industry has faced during the pandemic, the efforts made to attract talent to the industry, and the industry’s reliance on human capital (King et al., 2021; Liu & DeMicco, 2021; Popa et al., 2023). To address this gap, the current study uses control theory (Carver, 2006; Carver & Scheier, 2000) as a framework to examine how job search constraints associated with the COVID-19 pandemic can relate to job search stress among active job seekers in the hospitality industry, which then subsequently influences industry turnover intentions. Control theory suggests that individuals will self-regulate their job search efforts, continuously seeking to reduce a discrepancy between their current situation and a reference value. As suggested by control theory, an increase of COVID-19 job search constraints can lead to more job search stress, which will motivate applicants to reduce the discrepancy between their current situation and their career goals. One way to accomplish this is to seek employment outside the hospitality industry (King et al., 2021; Popa et al., 2023). In other words, if applicants feel that they are overly constrained and stressed from achieving their career goals in the hospitality industry, they may end up deciding to pursue a career in a different industry instead. * Corresponding author. University of Houston, Hilton College, 229 C. N. Hilton Hotel & College Houston, Texas, 77204-3028, United States. E-mail address: [email protected] (J.M. Madera). Contents lists available at ScienceDirect Journal of Hospitality and Tourism Management journal homepage: www.elsevier.com/locate/jhtm https://doi.org/10.1016/j.jhtm.2023.10.013 Received 31 January 2023; Received in revised form 12 September 2023; Accepted 20 October 2023

Journal of Hospitality and Tourism Management 57 (2023) 250–257 251 Carver and Scheier (2012) further argued that individual differences influence how individuals react to perceived stress. Therefore, we examined resilient coping as a moderator of the relationship between job search stress and industry turnover intentions. Resilient coping is the ability to recover or bounce-back during times of stress or adversity (Sinclair & Wallston, 2004). Grounded in Lazarus and Folkman’s (1984) framework, resilient coping includes efforts to manage demands appraised as stressful—that is, demands that exceed an individual’s resources. As an individual difference, some employees are better at coping during times of stress than others. Given the negative impact that the pandemic has had on the workplace, resilient coping might play an important role in how job seekers appraise job search stress related to COVID-19 job search constraints. The current study advances the literature in several ways. First, understanding the job search process from applicants’ perspective is an understudied area in hospitality literature (Madera et al., 2017; Tracey, 2014). This study seeks to address this gap by providing insight into hospitality applicants’ perspectives during the job application process and ways that companies could try addressing concerns and constraints related to the pandemic. Second, the COVID-19 pandemic is used to contextualize the current study given the fact that the widespread labor shortage facing the hospitality industry is an ongoing challenge. In August of 2021, for instance, a new record was set for the number of people quitting their jobs nationwide (Croes et al., 2021). That very same month, the rate of hospitality employees quitting their jobs was twice as high as the national average (Croes et al., 2021). Staffing continues to present a critical challenge for the hospitality industry in 2022, with a recent survey by the American Hotel & Lodging Association finding that 97% of their respondents were still experiencing staffing shortages (American HotelAssociation, 2022). Furthermore, as of June 2022, there were still approximately 16% fewer hotel/resort workers nationwide as compared to June 2019 (U.S. Bureau of Labor Statistics, 2022). Such labor shortages have impacted the hospitality industry deeply despite significant growth in demand since the onset of the COVID-19 pandemic (Maduro, 2022). Notable examples include hotels and resorts offering reduced services, and airports struggling with widespread delays and cancellations (Maduro, 2022). Consequently, this study has the potential for significant industry impact. The dilemma at hand is a major hurdle for hospitality companies, with critical impact in terms of profitability and maintaining customer satisfaction amid returning demand (Liu & DeMicco, 2021). The current paper, however, also provides valuable insights for organizations and policymakers to support job seekers, not only during crises like the pandemic but also in regular circumstances, where job search stress and industry transitions remain significant factors. Specifically, it advances job search research by exploring how individuals coping with high job search stress turn to alternative industries for their careers (van Hooft et al., 2021). While existing literature primarily focuses on finding similar roles in different companies within the same industry, it overlooks the possibility of individuals completely leaving their current industry due to job search constraints and stress. This integration bridges the job search and career turnover research literatures (Blau, 2007). 2. Theoretical development and hypotheses 2.1. Job search constraints during the pandemic: A control theory perspective The job search process has been described as a self-regulatory process in which job applicants monitor their progress and measure their progress toward a goal, such as attaining employment (Kanfer et al., 2001). Specifically, the job search process involves applicants making decisions regarding how much effort to put into applying to jobs, monitoring their progress, and appraising how close they are to fulfilling their goal of employment. As such, the job search process is highly dependent on the efforts of the job seeker. However, the job search process also involves external influences that can help or constrain their job seeking progress, such as job insecurity from the job market or work, financial limitations, personal/family problems that distract from job search efforts, failure to find employment, and/or setbacks in finding ideal organizations, making the job search process stressful (Wanberg et al., 2010; van Hooft et al., 2021). Control theory (Carver, 2006; Carver & Scheier, 2000) provides a theoretical lens to understand the role of self-regulation in the job search process. Control theory is based on a self-regulation model comprised of four components: perception, comparators, behaviors, and impact on environment. These components work together in a feedback loop to reduce perceived discrepancies between an individual’s current state and a reference value, such as attaining employment. The first component, perception, is the sensing of information about one’s current situation. Comparators, meanwhile, are a mechanism through which one compares one’s perception to a reference value, such as the goal of employment. When a discrepancy is perceived between the current state and the reference value, behaviors are taken. These are mechanisms used to reduce perceived discrepancies by impacting one’s environment or circumstances. This, in turn, will lead to a different perception and subsequently a reduced discrepancy between one’s current state and one’s reference value. In other words, control theory suggests that individuals will continuously seek to reduce the discrepancy between their current situation and reference value (Carver, 2006; Carver & Scheier, 1982, 2000). One major source for discrepancies in the job search process is job search constraints. These include child-care, family responsibilities, schedule conflicts, transportation restrictions, and financial constraints on job search-related items, such as professional clothing (Wanberg et al., 1999). While these everyday constraints can already make the job search experience a stressful one (Wanberg et al., 1999), the COVID-19 pandemic intensified many of these constraints, and even created additional obstacles. For example, the COVID-19 pandemic burdened many job seekers with increased child-care, family, and community responsibilities, as well as greater financial constraints (Koopmann et al., 2021). If job seekers happen to have immune-compromised family members or if they happen to be at a higher risk of contracting COVID-19 themselves, this could also be a significant new constraint in pursuing a job in the hospitality industry. Therefore, the pandemic likely had a disproportionate impact not only on those who were employed in the hospitality industry, but those who were seeking a job in the industry, by creating greater job search challenges, uncertainty, and job insecurity. Per control theory (Carver, 2006; Carver & Scheier, 1982, 2000), job seekers use their affect or perceptions, such as their perceived stress, as a way to assess their goal progress, subsequently affecting their efforts. Therefore, job search stress might be an important outcome of COVID-19 job search constraints. This suggests that the more COVID-19 job search constraints job seekers experience, the more likely they are to feel that they are failing at their goal of attaining employment, thereby increasing their job search stress. Job search stress involves feelings of anxiety and tension related to their job search efforts (Wanberg, 1997). Thus, the job search constraints predicated by the COVID-19 pandemic, such as financial troubles, feelings of job insecurity, and the uncertainty related to the pandemic, will be positively related to job search stress. This provides the basis for the first hypothesis of this study: H1. Greater COVID-19 job search constraints will lead to greater job search stress 2.2. Industry turnover intentions as an outcome of job search stress The term industry turnover intentions, also referred to as career change intentions, describes an employee’s intentions to leave their current occupation in pursuit of an occupation which would not be considered part of a normal career progression (McGinley et al., 2014). I. Popa and J.M. Madera

Journal of Hospitality and Tourism Management 57 (2023) 250–257 252 For instance, a front desk manager at a hotel leaving to pursue a new job as a technical support agent for an IT firm would be considered a career change. If an individual is likewise currently considering leaving their job for a position in a different industry, this would be considered industry turnover intentions, or career change intentions. Job search constraints may affect whether applicants look for alternative industry careers. These constraints sometimes impact job seekers during the job search process, decreasing job seekers’ perceptions of employability, that is, their likelihood of gaining employment or maintaining it (McGinley et al., 2020). In turn, past research has found support for a general decrease in perceptions of employability correlating with an increase in turnover intentions. Similarly, job insecurity – threats to a job, employment, or job quality – has been found to positively relate to turnover intentions (McGinley & Mattila, 2020), which could explain why even current hospitality industry employees might be choosing to leave their jobs in the industry. Past research indicates that turnover intentions could result from job-related stress. For instance, according to McGinley et al. (2014), dissatisfaction with career progression plays a key role in turnover intentions, as does work to life conflict. If applicants feel that they are overly constrained from achieving their career goals in the hospitality industry, they may end up deciding to pursue a career in a different industry instead. In accordance with control theory (Carver, 2006; Carver & Scheier, 2000), this tendency should increase along with COVID-19 constraints and subsequently experienced job search stress as applicants seek to reduce the discrepancy between their current situation and their career goals. This yields the following hypothesis: H2. Job search stress will mediate the relationship between COVID-19 job search constraints and intentions to leave the industry. 2.3. Resilient coping as a boundary condition Resilient coping is the ability to recover or bounce-back from stress or adversity caused by various factors, ranging from daily hassles at work to major life events (Sinclair & Wallston, 2004). As a personality trait, resilient coping has been conceptualized as a stable trait in which some individuals cope better than others during stressful times. In addition, resilient ‘coping has been examined as a trait that influences how people perceive their environment, such that it influences how an event is appraised (e.g., stressful) and how to cope with such appraisal of stress (Fletcher & Sarkar, 2013). For example, resilient coping has been shown to influence how employees perceive work events and environments, such that it affects employees’ work-related attitudes like their job satisfaction, career satisfaction, and organizational satisfaction (Hartmann et al., 2020). Thus, resilient coping is an important moderator of the relationship between stressful appraisals and work-related outcomes—employees with higher levels of resilient coping tend to be less negatively affected by stress at work, such as job demands or negative experiences. Although not examined within the confines of control theory, Carver and Scheier (2012) argued that individual differences or personality traits can influence how individuals react to perceived stress, thereby, placing resilient coping as a potential moderator of the relationship between job search stress and industry turnover intentions. Control theory (Carver & Scheier, 2012) points to the importance of stress appraisal via Lazarus and Folkman’s (1984) framework, such that resilient coping affects how one appraises stress or demands at work. In short, some job seekers will appraise job search stress related to COVID-19 job search constraints as less stressful than others, thereby positively influencing their ability to cope with such stress. Due to better stress coping abilities, job seekers with higher levels of resilient coping with be less likely to leave the industry in response to COVID-19-related job search stress and constraints. In contrast, job seekers with less resilient coping will be more likely to seek employment outside of the industry as a way to cope with the job search stress related to COVID-19 job search constraints. This is because control theory (Carver & Scheier, 2012) suggests that job seekers who appraise too much job search stress will be motivated to reduce the discrepancy between their current situation (i.e., too many job search constraints and their career goals of attaining employment); and one way to accomplish this is to leave the hospitality industry (King et al., 2021). In other words, if job seekers feel that they are overly constrained from achieving their career goals in the hospitality industry, they may end up deciding to pursue a career in a different industry instead. Thus, the relationship between job search stress and intentions to leave the industry will be stronger at lower levels of resilient coping (see Fig. 1 for the conceptual model). H3. Resilient cooping will moderate the relationship between job search stress and intentions to leave the industry, such that the relationship between job search stress and intentions to leave the industry will be stronger at lower levels of resilient coping. 3. Methodology 3.1. Design and procedure Given that control theory (Carver & Scheier, 2012) suggests that job search is a dynamic process, the current study used an experience sampling method (ESM) as the data collection strategy to examine the within-person relationships from the conceptual model. This strategy provides nested data at the within-person (Level 1) and between-person (Level 2) levels. The study collected data from the same participants in three waves using an event-contingent procedure (Yu et al., 2021). This procedure involved completing measures every two weeks over a six-week period, resulting in three time points of data. The participants were active job seekers, and this approach was chosen to capture sufficient job search experiences and variability in within-person measures. The selected timeframe was supported by previous studies (Yu et al., 2021; da Motta Veiga & Gabriel, 2016; Gabriel et al., 2021; Madera, 2018). The Level 1 variables include the measures of COVID-19 job search constraints, job search stress, and industry turnover intentions, which were completed at all three time points. These Level 1 variables are nested within Level 2 variables, which include the person-level variables, including the moderator resilient coping, and the demographic measures that were measured once at time 3. The data was collected in 2021.1 3.2. Sample A total of 282 responses were collected from the 95 active job seekers (males = 41%; females = 59%). The sample consisted of 95 active job seekers who were registered to attend a hospitality career fair at a university located in a major city in the southern region of the United States, satisfying the minimum recommended size of 50 at the betweenperson (Level 2) level (Yu et al., 2021; Heck et al., 2013). They reported an average of 3.2 h spent weekly devoted to job search activities. The participants had an average age of 24.95 (SD = 6.17) and an average tenure of 5.02 (SD = 4.52) years working in the hospitality industry. They identified themselves as 22.4% Caucasian, 5.3% as African 1 According to the Pew Research Center (2022), the work environment for employees in 2021 was still significantly impacted by the ongoing COVID-19 pandemic. Organizations were implementing health and safety protocols to ensure employee well-being, including measures such as physical distancing, mandatory mask-wearing, enhanced sanitation practices, and the provision of remote work support and resources. There was also a significant shift towards remote policies, such as reliance on technology for virtual hiring processes and an atmosphere of uncertainty in the job market. I. Popa and J.M. Madera

Journal of Hospitality and Tourism Management 57 (2023) 250–257 253 American/Black, 21.1% as Latino(a), 40.8% as Asian/Asian American, and 10.5% other.2 The participants all had an undergraduate college degree and therefore they were on the job market for professional roles, such as revenue management, sales, marketing, and finance/accounting. None were seeking entry-level or frontline positions. 3.3. Measures COVID-19 Job Search Constraints. The 4-item measure by Gabriel et al. (2021) was used to measure job search constraints related to COVID-19 using a 5-point Likert-type scale (1 = none at all to 5 = a great deal). The participants were instructed to report the extent to which the COVID-19 pandemic has interfered with their job search, over the last two weeks. Example items include: “How much has the COVID-19 pandemic in general interfered with your ability to look for a job?” and “How much have financial difficulties during the COVID-19 pandemic interfered with your ability to look for a job?”. The reliability for this measure was 0.88. Job Search Stress. We used the 4-item scale by Koopman et al. (2021) to measure job search stress using a 5-point Likert-type scale (1 = strongly disagree to 5 = strongly agree). The participants were instructed to indicate the extent to which they experienced stress while job searching over the last two weeks. Example items include: “I found myself getting agitated during my job search” and “I found it difficult to relax during my job search.” The reliability for this measure was 0.93. Industry Turnover Intentions. The 3-item measure by Yu, Lee, Popa, and Madera (2021) was used to measure industry turnover intentions using a 5-point Likert-type scale (1 = strongly disagree to 5 = strongly agree). The participants were instructed to report the extent to which they felt they should leave the industry: “I am actively searching for an alternative to the hospitality industry” and “As soon as it is possible, I will leave the hospitality industry.” The reliability for this measure was 0.90. Resilient Coping. We used the 4-item scale by Sinclair and Wallston (2004) to capture an individual’s ability to cope with stress in highly adaptive ways. The participants were instructed to report the extent to which they engage in coping behaviors using a 5-point Likert-type scale (1 = does not describe me at all to 5 = describes me very well). Example items include: “Regardless of what happens to me, I believe I can control my reaction to it” and “I look for creative ways to alter difficult situations.” The reliability for this measure was 0.79. Control variables. There were three control variables: week/time, weekly negative affect, and tenure in the industry. First, at Level 1, we controlled for the week/time to account for any spurious effects due to time (e.g., fatigue) that can manifest with repeated measures designs, such as ESM (Bolger & Laurenceau, 2013). Second, at Level 1, we controlled for bi-weekly negative affect to help mitigate common method bias (see Podsakoff et al., 2003) and to account for the role that negative affect has on active job searching (see da Motta Veiga et al., 2020). Watson et al.’s (1998) PANAS scale was used to measure negative affect with a 5-point Likert-type scale (1 = not at all to 5 = very much). The participants were asked to indicate the extent to which the felt negative emotions over the last two weeks of job searching. The reliability for this measure was 0.91. Third and last, at Level 2, we controlled for industry tenure given that people are less likely to transition into other industries and/or careers with longer tenure (Bedeian et al., 1992). 4. Results 4.1. Psychometric analyses Given the nest structure of the data, a multilevel confirmatory factor analysis (CFA) with two levels (Level 1 and Level 2) was used to analyze the construct validity of the measures (Dyer et al., 2005). Level 1 variables were centered at the within-person level and Level 2 variables were centered using the grand mean (Enders & Tofighi, 2007). The multilevel CFA showed good fit at the within- and between-person levels (χ2 = 155.22, df = 82, CFI = 0.98, TLI = 0.98, RMSEA = 0.049), suggesting a sufficient construct validity at both within-person and between-person level. We used the average variance extracted (AVE) scores and factor loadings to examine the convergence and discriminant validities. As shown in Table 1, the factor loadings ranged between 0.71 and 0.91, At the within-person level and at the between-person level the factor loadings ranged between 0.72 and 0.95. The AVEs ranged between 0.64 and 0.76 at the within-person level, and between 0.70 and 0.84 at the between-person level; all AVEs were above the 0.50 threshold, suggesting convergent validity at the within-person level (Kline, 2015). Adequate discriminant validity at both levels were found given that the AVEs of each construct at the within- and between-person levels were higher than the squared correlation coefficients as shown in Table 2. 4.2. Test of hypotheses We performed multilevel mediation analysis using Rockwood and Hayes’ (2017) MLmed macro for SPSS, which accounts for both the Level 1 and Level 2 variance, estimating the model parameters simultaneously (Hayes & Rockwood, 2020) to examine the direct and indirect effects (Hayes, 2017). As shown in Fig. 1, the analyzed model is a “1-1-1” multilevel mediation model (see Yu et al., 2021) with a Level 2 moderator given that the Level 1 variables (i.e., the repeated measures of COVID-19 job search constraints, job search stress, and industry turnover intentions) are nested within Level 2 variables (i.e., person-level variables, including the moderator: resilient coping). Monte Carlo confidence intervals, set at 10,000 samples, were calculated for the indirect effects and the index of moderated mediation. Null models testing the within-person variance intraclass correlation coefficient (ICC) showed significant within-person variability: 68% for COVID-19 job search Fig. 1. Conceptual model. 2 No significant differences by race/ethnicity emerged among the main variables (COVID19 job search constraints, job search stress, and industry turnover intentions), F(3, 73) = 0.73, p = 0.54; Wilk’s lambda = 0.97. I. Popa and J.M. Madera

Journal of Hospitality and Tourism Management 57 (2023) 250–257 254 constraints, 54% for job search stress; and 66% for industry turnover intentions. Table 3 shows the 1-1-1 within-person mediation model with a Level 2 moderator, controlling for day/time, weekly negative affect, and tenure in the industry. As shown in Table 2, the participants felt more job search stress during weeks in which they reported more COVID-19 job search constraints when compared to weeks in which they reported less COVID-19 job search constraints (γ = 0.38; SE = 0.08; CI.95 = 0.22, 0.53), thereby supporting Hypothesis 1. Additionally, participants reported higher intention to leave the industry on weeks they felt more job search stress compared to weeks they felt less job search stress (γ = 0.87; SE = 0.24; CI.95 = 0.38, 1.35). These effects were not significant at the between-person levels, suggesting that their weekly experiences with their job search mattered more for their intentions to leave the industry than their overall levels of COVID-19 job search constraints and job search stress across the three time points. The direct effect of COVID-19 job search constraints on intentions to leave the industry was not significant (γ = 0.14; SE = 0.08; CI.95 = − 0.03, 0.30). However, the indirect (i.e., mediation) effect of job search stress between COVID-19 job search constraints and intention to leave the industry was significant (γ = 0.33; SE = 0.12; CI.95 = 0.13, 0.58), thereby supporting Hypothesis 2. Lastly, the within-index of moderated mediation was significant, suggesting that resilient coping moderated the second stage of the mediation relationship between job search stress and intention to leave the industry (γ = − 0.8; CI.95 = − 0.14, − 0.03). A simple slope test (using one standard deviation below and above the mean) revealed that the positive relationship between job search stress and intention to leave the industry was weaker for job seekers with higher (simple slope = 0.18, p = 0.12) levels of resilient coping than lower levels (simple slope = 0.50, p = 0.01), thereby supporting Hypothesis 3. Table 1 Average variance extracted (AVE) scores, factor loadings, and composite reliabilities. Within-person level Between-person level Factor loading AVE Composite Reliability Factor loading AVE Composite Reliability COVID-19 Job Search Constraints 0.64 0.88 0.83 0.90 1. How much has the COVID-19 pandemic in general interfered with your ability to look for a job? 0.71 0.72 2. How much have your family responsibilities during the COVID-19 pandemic interfered with your ability to search for a job? 0.85 0.88 3. How much have financial difficulties during the COVID-19 pandemic interfered with your ability to look for a job? 0.85 0.89 4. How much have other responsibilities during the COVID-19 pandemic (e.g., internships, school, work, travel) interfered with your ability to search for a job? 0.80 0.85 Job Search stress 0.76 0.92 0.84 0.95 1. I found myself getting agitated during my job search. 0.83 0.88 2. I found it difficult to relax during my job search. 0.91 0.93 3. I felt that I was using a lot of nervous energy during my job search. 0.85 0.91 4. I found it hard to wind down during my job search. 0.89 0.95 Industry Turnover Intentions 0.76 0.90 0.70 0.94 1. I think a lot about leaving the hospitality industry. 0.81 0.87 2. I am actively searching for an alternative to the hospitality industry. 0.91 0.95 3. As soon as it is possible, I will leave the hospitality industry. 0.88 0.91 Table 2 Correlations and squared correlations. Mean Within SD Between SD 1 2 3 4 5 6 1. COVID-19 job search constraints 2.57 1.13 1.01 1 0.43* 0.27* 0.22 − 0.23* 0.42* 2. Job search stress 3.15 1.07 0.90 0.43* 1 .32* 0.27 − 0.14 0.52* 3. Industry turnover intentions 2.65 1.26 1.09 0.28* 0.33* 1 − 0.02 0.10 0.34* 4. Resilient coping 3.61 0.89 0.89 0.14 0.02 0.02 1 0.05 − 0.08 5. Industry tenure 5.07 4.52 4.53 − 0.12 − 0.08 0.04 0.15 1 − 0.19 6. Weekly NA 2.52 1.09 0.94 0.44* 0.59* 0.37* − 0.07 − 0.17 1 Note. Between-person (Level 2) correlations are shown in the bottom, left side. Within-person (Level 1) correlations are shown in the top, right side. *p<0.05. Table 3 Multilevel mediation model. Coefficient SE t p Job search distress as outcome Within-person effect COVID-19 job search constraints 0.38 0.08 4.67 0.001 Between-person effect COVID-19 job search constraints − 0.03 0.35 − 0.09 0.931 Industry turnover as outcome Within-person effect COVID-19 job search constraints 0.14 0.08 1.66 0.098 Job search stress 0.87 0.24 3.55 0.001 Interaction effect − 0.20 0.06 − 3.24 0.002 Between-person effect COVID-19 job search constraints 0.23 0.15 0.15 0.881 Job search stress 0.39 0.71 0.56 0.580 Resilient coping 0.13 0.63 0.21 0.836 Interaction effect − 0.05 0.19 0.21 0.805 Indirect Effect Effect SE MCLL MLUL Within-person effect 0.33 0.12 0.13 0.58 COVID-19 job search constraints → Job search stress → Industry turnover intentions Between-person effect 0.08 0.16 − 0.23 0.46 COVID-19 job search constraints → Job search stress → Industry turnover intentions Index of Moderated Mediation Estimate MCLL MLUL Within-person effect Resilient coping − 0.08 − 0.14 − 0.03 Between-person effect Resilient coping − 0.01 − 0.11 0.08 Note. Interaction effect = resilient coping × job search distress. I. Popa and J.M. Madera

Journal of Hospitality and Tourism Management 57 (2023) 250–257 255 5. Discussion 5.1. Theoretical implications The current study provides several theoretical implications. The findings of this study highlight that the impact of pandemic-related job search constraints on job search stress and intentions to leave the industry is a dynamic process rather than a static effect. The study focused on active job seekers and found that they experienced higher levels of job search stress during weeks when they faced more COVID-19 job search constraints compared to weeks with fewer constraints. Furthermore, job seekers reported greater intentions to leave the industry during weeks when they experienced higher job search stress compared to weeks with lower stress levels. Importantly, these effects were observed at the within-person level, indicating that job seekers’ weekly experiences with job search constraints and stress had a stronger influence on their intentions to leave the industry than their overall levels of COVID-19 job search constraints and stress across the entire study period. In other words, job seekers’ intentions were influenced by their week-to-week experiences and progress in their job search. These findings suggest that job seekers’ experiences and intentions can evolve over time, depending on their ongoing job search journey. It emphasizes the significance of collecting data over multiple time points rather than relying solely on cross-sectional data collection. This longitudinal perspective provides valuable insights into the dynamic nature of job search processes and underscores the need to consider the temporal aspects when studying job search outcomes (Yu, Lee, & Madera, 2021). Second, the current study found that COVID-19 job search constraints did not have a direct effect on intentions to leave the industry. Instead, COVID-19 job search constraints indirectly influence job seekers’ intentions to leave the hospitality industry through increasing their level of job search stress. This finding provides insight into the underlying mechanism through which COVID-19 job search constraints can have a negative impact on job seekers in the hospitality industry. It aligns with the principles of control theory, which suggests that when individuals perceive significant obstacles in achieving their career goals in a specific industry, such as the hospitality sector, they may choose to pursue opportunities in a different industry instead. According to control theory, this tendency is likely to increase as job seekers face higher levels of COVID-19 constraints and experience greater job search stress, as they try to minimize the discrepancy between their current circumstances and their desired career goals. By understanding this mechanism, organizations and policymakers can develop strategies to support job seekers and mitigate the negative effects of COVID-19 job search constraints. This may involve providing resources, guidance, and assistance to help job seekers cope with job search stress and overcome the challenges posed by the pandemic. Instances like the COVID-19 pandemic, however, are not the sole instances of adverse work events affecting industries disparately. Consider 9/11, which had a notably more detrimental effect on industries like aviation, tourism, and insurance (Makinen, 2002). Similarly, the Great Recession had a pronounced impact on sectors such as financial services and construction compared to others (Barello, 2014). These instances emphasize the necessity of studying how employees respond to such adverse industry-related events. These events, being inherently stressful and emotionally charged, may potentially give rise to unfavorable attitudes and behaviors towards their respective industries. Third, the current study also provides insight into how active job seekers cope with job search stress. Specifically, the current study revealed that resilient coping is an important moderator of the relationship between stressful appraisals of the job search process and intentions to leave the industry. By being able to cope better with job search stress, job seekers with higher levels of resilient coping were less likely to indicate plans to leave the industry. In contrast, when job seekers lower levels of resilient coping indicated higher levels of job search stress, they also reported higher intentions to pursue a career in a different industry. Thus, the relationship between job search stress and intentions to leave the industry was stronger at lower levels of resilient coping. Theoretically, this finding aligns with the viewpoint that coping is a process (Lazarus & Folkman, 1984) in which individual differences, such as resilient coping interact with environmental factors, like job search stress, that results from COVID-19 related job search constraints. When job seekers feel more job search stress, their resilient coping can weaken its influence on industry turnover. This finding is also important for the hospitality industry literature, given the labor shortage the industry has faced due to its heavy reliance on human capital (King et al., 2021; Liu & DeMicco, 2021). Lastly, the current study advances the job search literature by integrating it with the career/occupational turnover literature (Blau, 2007). To this end, the study at hand examines the phenomenon of seeking careers in alternative industries as an outcome of experiencing high levels of job search stress. Specifically, the job search literature has mainly focused on alternative work or moving from organization to organization working similar jobs (van Hooft et al., 2021), neglecting the fact that talent might leave the industry they have worked in (i.e., career/occupational turnover) when facing job search constraints and stress. Thus, this paper offers valuable insights not only for crisis situations like the pandemic but also for regular circumstances where job search stress and industry shifts are relevant. It enhances the job search literature by exploring how individuals dealing with high job search stress consider switching to entirely different industries. Previous job search studies mostly focused on changing roles within the same field, overlooking the possibility of individuals leaving their current industry altogether, which is also known as career/occupational turnover (van Hooft et al., 2021). In this manner, we bring together job search research and career/occupational turnover literature (Blau, 2007). 5.2. Practical implications The first practical implication for this study is gaining a better understanding of how job seekers make decisions to pursue employment in a different industry, which supports the idea of control theory. The findings suggest that job seekers regularly assess their progress towards their goal of securing a job in the hospitality industry throughout the job search process. If they perceive themselves as being far from achieving this goal, they may opt to explore job opportunities in alternative industries instead. To address this, it is important for hiring managers to effectively communicate with job applicants. This involves establishing clear expectations early in the application process, such as providing information on the expected timeframe for receiving a response regarding their application, and then adhering to that timeline. If additional interviews are necessary, it is crucial to clearly communicate this to the candidates as well. By maintaining clear and timely communication, hiring managers can help job seekers stay engaged and informed about their application status, reducing the likelihood of them pursuing opportunities in other industries due to perceived distance from their desired hospitality industry job. Another practical implication of this study is the recognition that job seekers may be deterred from pursuing hospitality jobs due to increased job search constraints caused by the pandemic. These constraints can encompass various factors such as technological accessibility issues associated with virtual applications and interviews, health concerns for both the job seeker and their immunocompromised family members, and added burdens related to childcare or personal responsibilities. In response to these significant challenges, hiring managers should proactively consider strategies to reduce barriers for potential candidates. This could involve implementing practices such as ensuring that the application process is user-friendly and easily understandable, offering flexible hours or scheduling options to accommodate individual needs, and effectively communicating the organization’s safety measures and protocols to job seekers who may have health-related concerns about I. Popa and J.M. Madera

Journal of Hospitality and Tourism Management 57 (2023) 250–257 256 working in the hospitality industry during pandemic conditions. By adopting these practices, hiring managers can create a more inclusive and supportive environment for job seekers, mitigating the impact of pandemic-related job search constraints and increasing the likelihood of attracting and retaining talented individuals to the hospitality sector. Lastly, our data shows the importance of increasing prospective job seekers’ resilience coping. Research shows that mindfulness training can help employees reduce stress levels by developing a heightened awareness of one’s experiences (Stacey & Cook, 2019). Specifically, mindfulness can enhance self-regulation skills and reframe one’s thinking as coping strategies for dealing stress, such as job search stress. Relying on a supportive network has also been found to increase employee resilience as it provides employees with a sense of belonging and exposure to alternative ideas and strategies to deal with constraints (Steinberg et al., 2016). 5.3. Limitations and future research As with all research, this study is not without limitations. The first limitation comes from the fact that this study used only one sample group. It is possible that individuals within this group, while a highly relevant demographic for the context of this study, may share certain similarities. For this reason, we recommend that future related studies use multiple samples, so as to compare results between potentially different demographics and increase the generalizability of the findings. A second limitation of this study is the generalizability of the findings based on context. The COVID-19 pandemic and related job search constraints were the context for this particular study, and while this is a significant and impactful global event, future research should look into replicating the findings in different contexts. This is important to rule out the possibility of unique features of the COVID-19 pandemic resulting in the specific results observed, thereby further improving generalizability. A qualitative approach can provide a deeper understanding of how the pandemic affected not only the participants’ job search process but also their perceptions and expectations regarding the hospitality industry as a viable career path. For instance, the pandemic may have intensified feelings of job insecurity, a prevalent type of job search constraint (Abbas et al., 2021). While our current study used a quantitative approach and assessed general COVID-19-related job search constraints (Gabriel et al., 2021), it was unable to capture specific types of constraints caused by the pandemic. To address this limitation, future research should employ a qualitative approach to explore these specific search constraints. Future research might examine how cultural values can influence the relationships examined in the current study. For example, research shows that cultural values can influence how employees perceive and respond to job insecurity such that employees with a stronger individualistic orientation tend to experience more negative reactions to job insecurity, such as decreased job satisfaction and increased turnover intentions than employees with more collectivist orientation (Probst & Lawler, 2006). Thus, employees from countries that emphasize individualism might experiences more job search stress due to COVID-19 related job search constraints. Lastly, it is important to note that the current study focused on one specific moderator, namely job search stress, in the relationship between job search stress and intentions to leave the industry. However, there are other factors that could potentially influence this relationship. Previous research has highlighted the role of financial needs as a motivator for job seeking behavior. Studies have found that the impact of negative workplace experiences, such as incivility, on job search behavior is more pronounced for employees with lower income levels (Megeirhi et al., 2020). This suggests that perceived financial need and financial situation can act as moderators in the relationship between job search stress and intentions to leave the industry. In other words, individuals who have a negative financial situation or a strong financial need may experience a stronger impact of job search stress on their job search behaviors. This is because the combination of high job search stress and financial strain can create a heightened sense of urgency and motivation to find alternative employment options (Boswell et al., 2012). The stress of the job search process may be amplified for individuals who are facing financial difficulties, leading to a stronger link between job search stress and intentions to leave the industry. Considering the moderating role of perceived financial need and financial situation can provide a more comprehensive understanding of the complex dynamics between job search stress, industry turnover intentions, and external factors. Future research should explore additional moderators to gain a more nuanced understanding of how various individual and situational factors interact to shape job search behavior and career decisions. Regardless of these potential limitations, the current study advances the job search literature by integrating the job search literature with the career/occupational turnover literature (Blau, 2007). Furthermore, although this research centered around the COVID-19 pandemic to examine how job constraints contribute to industry turnover intentions through job search stress, its findings hold broader implications. They provide important implications for organizations and policymakers, applicable not only in crises like the pandemic, but also in typical situations where job search stress and industry changes are pivotal. Specifically, the current study examined how job search constraints might motivate hospitality talent to leave the industry—an industry that heavily relies on human capital and is also facing a labor shortage (King et al., 2021; Liu & DeMicco, 2021). Using control theory (Carver, 2006; Carver & Scheier, 2000) as a framework, the current study found that job search constraints related to job search stress among active job seekers in the hospitality industry, which then subsequently influences industry turnover intentions. This relationship was strongest for job seekers with lower levels of resilient coping, suggesting that individual differences play an important role in how job seekers cope with job search constraints. Declaration of competing interest There are no conflict of interest or financial disclosures to report. References Abbas, M., Malik, M., & Sarwat, N. (2021). Consequences of job insecurity for hospitality workers amid COVID-19 pandemic: Does social support help? Journal of Hospitality Marketing & Management, 30(8), 957–981. American Hotel, & Association, L. (2022). As 97% of surveyed hotels report staffing and experience sampling research. Guilford press. Barello, S. H. (2014). Consumer spending and US employment from the 2007-2009 recession through 2022. Monthly Labor Review, 137, 1–32. Bedeian, A. G., Ferris, G. R., & Kacmar, K. M. (1992). Age, tenure, and job satisfaction: A tale. Blau, G. (2007). Does a corresponding set of variables for explaining voluntary organizational turnover transfer to explaining voluntary occupational turnover? Journal of Vocational Behavior, 70(1), 135–148. Bolger, N., & Laurenceau, J. P. (2013). Intensive longitudinal methods: An introduction to diary. Boswell, W. R., Zimmerman, R. D., & Swider, B. W. (2012). Employee job search: Toward an understanding of search context and search objectives. Journal of Management, 38 (1), 129–163. Carver, C. S. (2006). Approach, avoidance, and the selfregulation of affect and action. Motivation and Emotion, 30, 105–110. Carver, C. S., & Scheier, M. F. (1982). Control theory: A useful conceptual framework for chen, C., C., & chen, M. H. (2021). Well-Being and career change intention: COVID19’s impact on unemployed and furloughed hospitality workers. International Journal of Contemporary Hospitality Management, 33(8), 2500–2520. Carver, C. S., & Scheier, M. F. (2000). On the structure of behavioral self-regulation. In M. Boekaerts, P. R. Pintrich, & M. Zeidner (Eds.), Handbook of self-regulation: 41–84. San Diego: Academic. Carver, C. S., & Scheier, M. F. (2012). Attention and self-regulation: A control-theory approach to human behavior. Springer Science & Business Media. Chen, C. C., & Chen, M. H. (2021). Well-being and career change intention: COVID-19’s impact on unemployed and furloughed hospitality workers. International Journal of Contemporary Hospitality Management, 33(8), 2500–2520. Chen, H., & Eyoun, K. (2021). Do mindfulness and perceived organizational support work? Fear of COVID-19 on restaurant frontline employees’ job insecurity and I. Popa and J.M. Madera

Journal of Hospitality and Tourism Management 57 (2023) 250–257 257 emotional exhaustion. International Journal of Hospitality Management, 94, Article 102850. Croes, R. R., Semrad, K., & Rivera, M. A. (2021). The state of the hospitality industry 2021. Dyer, N. G., Hanges, P. J., & Hall, R. J. (2005). Applying multilevel confirmatory factor analysis techniques to the study of leadership. The Leadership Quarterly, 16(1), 149–167. Enders, C. K., & Tofighi, D. (2007). Centering predictor variables in cross-sectional multilevel models: A new look at an old issue. Psychological Methods, 12(2), 121–138. Fletcher, D., & Sarkar, M. (2013). Psychological resilience: A review and critique of definitions, concepts, and theory. European Psychologist, 18(1), 12. Gabriel, A. S., MacGowan, R. L., Ganster, M. L., & Slaughter, J. E. (2021). The influence of COVID-induced job search anxiety and conspiracy beliefs on job search effort: A within-person investigation. Journal of Applied Psychology, 106(5), 657. Guzzo, R. F., Wang, X., Madera, J. M., & Abbott, J. (2021). Organizational trust in times of COVID-19: Hospitality employees’ affective responses to managers’ communication. International Journal of Hospitality Management, 93, Article 102778. Han, H., Koo, B., Ariza-Montes, A., Lee, Y., & Kim, H. R. (2021). Are airline workers planning career turnover in a post-COVID-19 world? Assessing the impact of risk perception about virus infection and job instability. Journal of Hospitality and Tourism Management, 48, 460–467. Hartmann, S., Weiss, M., Newman, A., & Hoegl, M. (2020). Resilience in the workplace: A multilevel review and synthesis. Applied Psychology, 69(3), 913–959. Hayes, A. F. (2017). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. Guilford publications. Hayes, A. F., & Rockwood, N. J. (2020). Conditional process analysis: Concepts, computation, and advances in the modeling of the contingencies of mechanisms. American Behavioral Scientist, 64(1), 19–54. Heck, R. H., Thomas, S. L., & Tabata, L. N. (2013). Multilevel and longitudinal modeling with IBM SPSS. Routledge. https://restaurant.org/Covid19. van Hooft, E. A., Kammeyer-Mueller, J. D., Wanberg, C. R., Kanfer, R., & Basbug, G. (2021). Job search and employment success: A quantitative review and future research agenda. Journal of Applied Psychology, 106(5), 674. Kanfer, R., Wanberg, C. R., & Kantrowitz, T. M. (2001). Job search and employment: A personality–motivational analysis and meta-analytic review. Journal of Applied Psychology, 86(5), 837. King, C., Madera, J. M., Lee, L., Murillo, E., Baum, T., & Solnet, D. (2021). Reimagining attraction and retention of hospitality management talent–A multilevel identity perspective. Journal of Business Research, 136, 251–262. Koopmann, J., Liu, Y., Liang, Y., & Liu, S. (2021). Job search self-regulation during COVID-19: Linking search constraints, health concerns, and invulnerability to job search processes and outcomes. Journal of Applied Psychology, 106(7), 975. Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and coping. Springer publishing company. Liu, L., & DeMicco, F. (2021). Labor shortages and increasing labor costs post-covid-19: How future Hospitality businesses are going to thrive? Hospitality net. https://www.hospitalit ynet.org/opinion/4106002.html. Madera, J. M. (2018). When targets blame their organization for sexual harassment: A multilevel investigation of within-person appraisals. Cornell Hospitality Quarterly, 59 (1), 49–60. Madera, J. M., Dawson, M., Guchait, P., & Belarmino, A. M. (2017). Strategic human resources management research in hospitality and tourism. International Journal of Contemporary Hospitality Management, 29(1), 48–67. Maduro, F. (2022). Labor shortages in hospitality industry. LinkedIn. https://www.linkedin. com/pulse/labor-shortages-hospitality-industry-frank-maduro. Makinen, G. (2002). The economic effects of 9/11: A retrospective assessment. Washington, DC: Library of Congress. McGinley, S., & Mattila, A. S. (2020). Overcoming job insecurity: Examining grit as a predictor. Cornell Hospitality Quarterly, 61(2), 199–212. McGinley, S., Mattila, A. S., & Self, T. T. (2020). Deciding to stay: A study in hospitality managerial grit. Journal of Hospitality & Tourism Research, 44(5), 858–869. McGinley, S., O’Neill, J., Damaske, S., & Mattila, A. S. (2014). A grounded theory approach to developing a career change model in hospitality. International Journal of Hospitality Management, 38, 89–98. Megeirhi, H. A., Ribeiro, M. A., & Woosnam, K. M. (2020). Job search behavior explained through perceived tolerance for workplace incivility, cynicism and income level: A moderated mediation model. Journal of Hospitality and Tourism Management, 44, 88–97. da Motta Veiga, S. P., & Gabriel, A. S. (2016). The role of self-determined motivation in job search: A dynamic approach. Journal of Applied Psychology, 101(3), 350–361. da Motta Veiga, S. P., Sun, S., Turban, D. B., & Foo, M. D. (2020). How does affect relate to job search effort and success? It depends on pleasantness, activation, and core selfevaluations. Human Resource Management, 60(6), 921–933. National Restaurant Association. (2020). Coronavirus information and resources. of two perspectives. Journal of Vocational Behavior, 40(1), 33–48. https://doi.org/10.1037/ 0033-2909.92.1.111. personality-social, clinical, and health psychology. Psychological Bulletin, 92(1), 111–135. Pew Research Center. (2022). COVID-19 pandemic continues to reshape work in America. Retrieved from https://www.pewresearch.org/social-trends/wp-content/uploads/s ites/3/2022/02/PSDT_2.16.22_covid_work_report_clean.pdf. Podsakoff, P. M., MacKenzie, S. B., Lee, J.-Y., & Podsakoff, N. P. (2003). Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88(5), 879–903. Popa, I., Lee, L., Yu, H., & Madera, J. M. (2023). Losing talent due to COVID-19: The roles of anger and fear on industry turnover intentions. Journal of Hospitality and Tourism Management, 54, 119–127. Probst, T. M., & Lawler, J. (2006). Cultural values as moderators of employee reactions to job insecurity: The role of individualism and collectivism. Applied Psychology, 55(2), 234–254. shortages, AHLA Foundation expands recruitment campaign. AHLA https: //www.ahla.com/press-release/97-surveyed-hotels-report-staffing-shortages-ahla-f oundation-expands-recruitment. Sinclair, V. G., & Wallston, K. A. (2004). The development and psychometric evaluation of the brief resilient coping scale. Assessment, 11(1), 94–101. Stacey, G., & Cook, G. (2019). A scoping review exploring how the conceptualisation of resilience in nursing influences interventions aimed at increasing resilience. International Practice Development Journal, 9(1). Steinberg, B., Klatt, M., & Duchemin, A. (2016). Feasibility of a mindfulness-based intervention for surgical intensive care unit personnel. American Journal of Critical Care, 26(1), 10–18. https://doi.org/10.4037/ajcc2017444 Tracey, J. B. (2014). A review of human resources management research. International Journal of Contemporary Hospitality Management, 26(5), 679. U.S. Bureau of Labor Statistics. (2021). Covid-19 ends longest employment recovery and expansion in CES history, causing unprecedented job losses in 2020 : Monthly Labor Review. U.S. Bureau of Labor Statistics. https://www.bls.gov/opub/mlr/2021/articl e/covid-19-ends-longest-employment-expansion-in-ces-history.htm. U.S. Bureau of Labor Statistics. (2022). Current employment statistics - CES (national). U.S. Bureau of Labor Statistics. https://www.bls.gov/ces/. Wanberg, C. R. (1997). Antecedents and outcomes of coping behaviors among unemployed and reemployed individuals. Journal of Applied Psychology, 82(5), 731–744. Wanberg, C. R., Kanfer, R., & Rotundo, M. (1999). Unemployed individuals: Motives, jobsearch competencies, and job-search constraints as predictors of job seeking and reemployment. Journal of Applied Psychology, 84(6), 897–910. https://doi.org/ 10.1037/0021-9010.84.6.897 Wanberg, C. R., Zhu, J., & Van Hooft, E. A. J. (2010). The job-search grind: Perceived progress, self-reactions, and self-regulation of search effort. Academy of Management Journal, 53, 788–807. Yu, H., Lee, L., & Madera, J. M. (2021). Collecting repeated data over time: Applying experience sampling methodology to the hospitality management context. Cornell Hospitality Quarterly, 62(1), 62–75. Yu, H., Lee, L., Popa, I., & Madera, J. M. (2021). Should I leave this industry? The role of stress and negative emotions in response to an industry negative work event. International Journal of Hospitality Management, 94, Article 102843. I. Popa and J.M. Madera

Journal of Hospitality and Tourism Management 57 (2023) 61–71 Available online 10 September 2023 1447-6770/© 2023 The Authors. Published by Elsevier Ltd. on behalf of CAUTHE - COUNCIL FOR AUSTRALASIAN TOURISM AND HOSPITALITY EDUCATION. All rights reserved. Journaling memorable and meaningful tourism experiences: A strengths-based approach to technology-mediated reminiscence C.K. Bruce Wan a,* , Cees J.P.M. de Bont b , Paul Hekkert c , Sebastian Filep d , Kenny K.N. Chow e a School of Arts and Social Sciences, Hong Kong Metropolitan University, 30 Good Shepherd St., Ho Man Tin, Hong Kong b School of Design and Creative Arts, Loughborough University, Leicestershire, LE11 3TU, UK c Faculty of Industrial Design Engineering, TU Delft, Landbergstraat 15, 2628 CE, Delft, the Netherlands d School of Hotel and Tourism Management, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong e College of Communication, National Chengchi University, No. 64, Section 2, Zhinan Rd, Wenshan District, Taipei City, 11605, Taiwan ARTICLE INFO Keywords: Memorable and meaningful tourism experiences (MMEs) Reflective technology Character strengths Savouring Positive psychology interventions Travel diary ABSTRACT Reminiscing on memorable travel experiences is a common practice amongst many travellers. This study introduces positive psychology interventions – cultivation of character strengths and savouring strategies - to examine memorable and meaningful tourism experiences (MMEs). Although both interventions aim to increase well-being, little research has been conducted on their roles in enriching MMEs. MMEs are fundamental to understand as part of the travel reminiscence process. MMEs could be heightened by connecting tourists’ past experiences with their character strengths (capacities for ways of behaving). Savouring, on the other hand, facilitates the connections to places. The reminiscence process helps tourists gain self-knowledge and make wellbeing oriented choices in their future journeys. In so doing, this research study created an interactive strengthsbased journal that facilitated tourists to incorporate their character strengths in their past MMEs. The narratives were structured to connect explicit experiential components, such as tourism activities, with implicit psychological factors, such as emotions, character strengths, and values. Data collection involved ten tourists of diverse nationalities who created 51 MME narratives. Participants were then invited to savour their strengths used, reflect on their narratives, and express their behavioural intentions for their next trip. Data analysis, using grouped frequency distributions, found that MMEs were associated with the moderate strengths rather than the signature (prominent) strengths of the participants, such as curiosity and gratitude. Appreciation of beauty and excellence was the most dominant strength observed. The findings showed participants preferred their future journeys to be congruent with their character strengths. Theoretical and practical implications for tourist experience research are outlined. 1. Introduction Tourists commonly anticipate, savour and remember key moments of their travel journeys. Tourists who perceive tourist experiences as memorable and meaningful look forward to similar experiences in the future (Yan & Halpenny, 2022). Past research (Hosseini et al., 2021; Tung & Ritchie, 2011) has recognised these memorable and meaningful experiences (MMEs) as particularly valuable and gratifying for tourists. MMEs are a reliable indicator of tourist satisfaction (Otto & Ritchie, 1996) and lead to repeated purchases, revisitation and place attachment (Rejikumar et al., 2021). Emotions, sense-making, and memory processes are at the core of any individual tourist experience. Emotions, which are heightened in MMEs, enhance the vividness of episodic tourist experiences which are registered in the long-term memories of individuals (Levine & Pizarro, 2004) and lead to sense-making. Sense-making refers to the process by which tourists attribute meanings to episodic experiences, which become comprehensible, worthwhile, and significant to them (Steger, 2016). Nevertheless, even though tourists may be participating in the same activity in the same location at a destination, their experiences can vary greatly (Ooi, 2006). Therefore, exploring the inner psychological dimensions pertaining to MMEs can provide a fuller picture of tourists’ well-being and their subsequent behavioural intention. A seminal study conducted by Kim et al. (2012) defined the * Corresponding author. E-mail addresses: [email protected] (C.K.B. Wan), [email protected] (C.J.P.M. de Bont), [email protected] (P. Hekkert), sebastian.filep@ polyu.edu.hk (S. Filep), [email protected] (K.K.N. Chow). Contents lists available at ScienceDirect Journal of Hospitality and Tourism Management journal homepage: www.elsevier.com/locate/jhtm https://doi.org/10.1016/j.jhtm.2023.08.017 Received 20 September 2022; Received in revised form 21 August 2023; Accepted 27 August 2023

Journal of Hospitality and Tourism Management 57 (2023) 61–71 62 memorable tourism experience as a tourism experience positively remembered and recalled after the event has occurred. The study identified seven tourism experience factors: hedonism, refreshment, local culture, meaningfulness, knowledge, involvement, and novelty. Yet, the theorization of memorable tourism experiences remains weak because other studies failed to replicate the scale (Hosany et al., 2022). The discrepancy may be due to the multifaceted nature of memorable tourism experiences, the different theoretical frameworks and research instruments used. Related tourism and positive psychology research (Filep & Laing, 2018; Vada et al., 2020) has shown interest in hedonic and eudaimonic well-being, which involves tourism experiences that promote personal growth and a higher level of psychological functioning. In particular, recent studies (Miyakawa et al., 2022; Zhang, 2023) showed eudaimonic well-being is strengthened with savouring and strengths cultivation. The combination of the two interventions would allow tourists to gain deeper insights (e.g., self-realisation and self-acceptance) from their past MMEs which could lead to personal growth according to their personality traits and values (Carruthers & Hood, 2004). In other words, reminiscing about past MMEs and building one’s awareness on their strengths used in these MMEs could optimise the well-being effects. To further explore the complexity of tourist well-being, Hosany et al. (2022) suggested the need to examine the diversity and individual nature of MME experiences through cross-cultural settings, facilitating participants’ recall, and using multiple theoretical perspectives. This research study responds to these research gaps. It introduces positive psychology interventions — character strengths intervention (Niemiec, 2017) and the savouring intervention (Bryant & Veroff, 2007) to study MMEs. This study specifically aims to gain an in-depth understanding of MMEs by connecting them with the character strengths of the tourists in the post-trip savouring process. In so doing, the study develops a novel research instrument (an interactive journaling platform) to guide tourists to connect their character strengths with their past MMEs through reminiscing and reflection. Two research questions are addressed. First, how do tourists draw upon their character strengths in their MMEs? And second, how do tourists take their MMEs into consideration when planning future trips? This study provides new insights into reminiscence of tourist experiences by understanding their character strengths and the subsequent behavioral intention. It also informs the design of smart tourism platforms and informatics systems that promote reminiscing and encourage tourists to strive for self-improvement in their future trips. 2. Literature review 2.1. Positive psychology, character strengths and savouring Character strengths are “pre-existing capacities for a particular way of behaving, thinking, or feeling that are authentic and energizing to the user and enable optimal functioning, development, and performance” (Linley, 2008, p. 9). The Values in Action Classification of character strengths identified six core virtues and 24 related character strengths (Table 1) that contribute to human flourishing—the optimal continuing development of human potential. There are creativity, curiosity, judgement, love of learning, and the like (see Table 1). Although Biswas-Diener et al. (2011) argue most individuals possess all 24 character strengths to different degrees, signature strengths are the most prominent strengths that represent one’s authentic identity and capabilities. However, only about one-third of people can identify their own strengths, and fewer use them consciously in their lives (Biswas-Diener et al., 2011). The VIA-IS (https://www.viacharacter.org/) is a 240-question online survey that allows people to find their strengths in a ranking order. Strengths-based interventions are activities and exercises designed to help individuals identify and cultivate their unique strengths and virtues, aiming to improve overall well-being and lead a more fulfilling and meaningful life (Gander et al., 2013). With more than 70 interventions developed, most of them consist of a three-step process: the so called aware-explore-apply model (Niemiec, 2017). The awareness stage guides people to become aware of their own strengths, which were implicit to them, and then develop a comprehensive understanding of their signature strengths. The explore stage, which is similar to the savouring process, allows tourists to delve deeper into their strengths by connecting them with important moments of their lives (e.g., memorable and meaningful events). The apply stage involves seeking out new opportunities and devising an action plan to put their strengths to good use, for instance, making positive changes in their lives and others. Prior research studies (Gander et al., 2013) have found that consciously exercising one’s signature strengths significantly increases one’s life satisfaction and alleviates depression. Empirical studies on character strengths in tourism studies however are almost non-existent. Warren and Coghlan (2016) investigated what character strengths might encourage travellers to engage in on-site pro-environmental behaviours. A more recent study (Li et al., 2020) connects digital-free tourism with Table 1 Values in action classification of character strengths and virtues (Peterson & Seligman, 2004). Virtues Character strengths Virtues Character strengths Wisdom Creativity – originality, adaptive, ingenuity; Curiosity – interest, novelty-seeking, exploration, openness to experience; Judgment – critical thinking, thinking things through, openmined; Love of learning – mastering new skills & topics, systematically adding to knowledge; Perspective – wisdom, providing wise counsel, taking the big picture view. Transcendence Appreciation of beauty & excellence – awe, wonder, elevation; Gratitude – thankful for the good, expressing thanks, feeling blessed, Hope – optimism, future-mindedness, future orientation; Humor – playfulness, bringing smiles to others, light-hearted; Spirituality – religiousness, faith, purpose, meaning Courage Bravery – valor, not shrinking from fear, speaking up for what’s right; Perseverance – persistence, industry, finishing what one starts; Honesty – authenticity, integrity; Zest – vitality, enthusiasm, vigor, energy, feeling alive and activated. Temperance Forgiveness – mercy, accepting others’ shortcomings, giving people a second chance; Humility – modesty, letting one’s accomplishments speak for themselves; Prudence – careful, cautious, not taking undue risks; Self-regulation – selfcontrol, disciplined, managing impulses & emotions; Humanity Love – both loving and being loved, valuing close relations with others; Kindness – generosity, nurturance, care, compassion, altruism, “niceness”; Social intelligence – emotional intelligence, being aware of the motives/ feelings of oneself/ others, knowing what makes other people tick. Justice Teamwork – citizenship, social responsibility, loyalty; Fairness – just, not letting feelings bias decisions about others; Leadership – organizing group activities, encouraging a group to get things done. C.K.B. Wan et al.

Journal of Hospitality and Tourism Management 57 (2023) 61–71 63 specific character strengths, such as travellers’ self-regulation, social intelligence, and open-mindedness. The topic of savouring has equally received very little research attention. Savouring is “not just the awareness of pleasure but also a conscious attention to the experience of pleasure” (Bryant & Veroff, 2007, p. 12). The savouring intervention is a behaviour change strategy that cultivates people’s capabilities of appreciating positive experiences conducive to integrated self-development. In the context of tourism, savouring can take place before the trip (anticipation), during the trip (in the moment), and after the trip (reminiscence). Despite the positive effects that the savouring intervention has on people’s well-being, it is primarily used by therapists, coaches, and educators (Biswas-Diener et al., 2011). Only recently has savouring garnered researchers’ attention in tourism (Miyakawa et al., 2022). Prior savouring studies explored the post-travel reminiscence process and the potential for place attachment (Yan & Halpenny, 2021) but their role in MMEs is poorly understood. 2.2. The nature of MMEs and episodic memories MMEs are stored in the brain as episodic memories, composed of autobiographical and flashbulb memories (Skavronskaya et al., 2017). Autobiographical memory refers to the memory of personally experienced events and episodes. The flashbulb memory, on the other hand, is a type of highly vivid and emotionally charged memory that is triggered by a surprising and consequential news event (Brown & Kulik, 1977). Despite the vividness and high emotional involvement of flashbulb memories, these memories are not especially accurate (Talarico & Rubin, 2003). Although storytelling is an effective way to capture MMEs (Moscardo, 2017), the accuracy of the retrospective report will be limited to memory errors and the influence of recall bias (Yüksel, 2017). This study mitigates problems associated with retrospective storytelling. Given that digital photo is the most popular medium that travellers use to capture memorable and meaningful moments during their travel journeys (Mang et al., 2016), textual content allows travellers to record personal thoughts and gain insights from their experiences (Hiemstra, 2001). Structured journaling with photo elicitation would enable travellers to accurately reconstruct their MMEs, as diaries created by travellers are often inadequate for post trip reminiscing (Wan, 2019). An effective reminiscing process requires travellers to accurately recall factual and experiential dimensions of their experiences in the narratives, so that memorabilia (i.e., photos) can better support reflection and introspection. Therefore, the journal needs to provide a narrative structure (e.g., context, orientation, key moments, value gained) and experience retrieval cues (e.g., “What did you do?”, “Who was with you?”, “What did you feel?”, “What were the values gained?”) to facilitate travellers in narrating their MMEs, connecting their strengths with their experiences, and reminisce them (Hosany et al., 2022). Thus, building the strengths-based journal on an interactive digital platform can help the researchers to collect more accurate data from participants because technology allows for creating personalized strength profiles using user-generated content and generating personalized reports for reminiscing. 3. Research methodology To meet the aim and address the two research questions, a proof-ofconcept interactive journaling platform was created to guide participants in creating strengths-based narratives on their MMEs. The narratives were composed of both implicit psychological dimensions (i.e., emotion, character strengths, and values) and explicit experiential dimensions (i.e., activities, places, and people) of MMEs. 3.1. Features of the strengths-based journaling platform The journaling platform composed of four sections: user profile, story creation, story browser, and reflection. User Profile: User profile mainly consists of the strengths profile from the results of VIA-IS survey conducted prior to the account creation. A short description is provided on each strength which allows users to become familiar with the definition (Fig. 1(a)). Story creation: This section guides users to create MMEs entries (see Fig. 1(b) for an overview). It consists of nine subsections in which users can upload images, create tags, input descriptions, assign emotions, associate character strengths, provide titles and dates of their experience. The nine sections, which are created based on strengths-based interventions guide users to connect explicit tourism activities and implicit psychological dimensions. The nine sections include people involved, places visited, activities taken, peak moments experienced, character strengths used, the values gained (i.e., the significance of the experience), the title of the story, and the date of the experience. Here, three subsections helped elicit the implicit psychological dimensions of MMEs from users. First, to facilitate users’ expressions, an emotion dial (Fig. 1(d)) was created to allow users to select a representative emoticon to describe the emotions felt of their lived moments. This feature allowed users to express non-verbal dimensions of their travel experiences. Second, users were invited to select a maximum of two character strengths they had drawn upon in the experience. The strengths were arranged according to the user’s strength profile to increase usability (Fig. 1(e)). Users were allowed to choose multiple strengths of the experience because strengths are not expressed in isolation but in combination with one another (Biswas-Diener et al., 2011). For each of the strengths selected, users were invited to provide a short description of its contribution to the experience. The last subsection invited users to select the benefits gained as a result of the MMEs. The value section (Fig. 1(f)) listed 17 items which cover hedonic and eudaimonic well-being dimensions that may be associated with tourism activities. Since there is little consensus on what constitutes eudaimonic well-being (Biswas-Diener et al., 2009), this section aggregates well-being dimensions from prominent frameworks in positive psychology and tourism. These frameworks include Ryff’s model of well-being (Ryff, 1989), self-determination theory (Deci & Ryan, 1985), and memorable tourism experience scales (Chandralal & Valenzuela, 2015; Kim et al., 2012). The list includes self-acceptance, personal growth, meaning and purpose in life, sense of mastery, autonomy, kinship, friendship, health, better world and society, prosperity, wisdom, social recognition, harmony, excitement, knowledge, courage, and justice. Users can register a maximum of two values gained for each narrative. Story browser: Once the stories were created, they could be viewed in the story browser (Fig. 2(a)). This section allowed users to browse all stories created. Each story is displayed with a picture and a title. Clicking on the picture would bring users to the corresponding story (see Fig. 3 as an example). The filter feature (Fig. 2(b)) allowed users to review entries by selecting specific strengths, values, and people. Reflection: This section consisted of two components: a dashboard and series of insight cards. The dashboard summarised the strengths related information aggregated from all entries: an overview of the story and tag created; the most used strengths and the top valued gained (Fig. 2(c)). Clicking the elements on the dashboard would bring users to the story browser with the specific content. The purpose of the dashboard is to make the implicit dimension of MMEs explicit which users can inspect. Below the dashboard, a number of insights and suggestions derived from users’ entries are listed. The insight cards display users on five attributes in connection with their MMEs: the values gained (Fig. 2 (d)), the strengths involved (Fig. 2(e)), the relationships created (Fig. 2 (f)), the activities involved, and the place visited. Using this section allowed the researchers to gather users’ feedback on their future trip plans with these attributes in mind. By clicking the button on each of the insight cards, participants were asked whether they wanted to develop their strengths, pursue the same values, travel with the same travel mate (s), do the same activity, and visit the same place on their future C.K.B. Wan et al.

Journal of Hospitality and Tourism Management 57 (2023) 61–71 64 journeys (Fig. 2(g)). 3.2. Participant recruitment and data collection Data collection was conducted in four steps: recruiting participants, collecting their strength profiles, creating stories with the strengthsbased journaling platform and interviewing after finishing the tasks. This study used a purposive sampling, where each participant needed to provide at least five MMEs with a fair amount of photos captured on each of these experiences. To the best of the researchers’ knowledge, there is no optimal timeframe for reminiscence process to take place. However, extant research (Brown & Kulik, 1977) has found that novel, distinctive, emotionally charged and personally significant experiences can create long-term episodic and autobiographical memories which can last for 5–30 years. This study recruited ten participants through posters and social media platforms (Table 2) during the May–July period of the year 2020. A suitable participant had to have experienced at least five MMEs in the past five years. Owing to the distinctive nature of MMEs, participants self-proclaimed that the stories were memorable and meaningful. Each participant had to provide at least ten photos of every experience and then discuss these experiences. Among ten participants, seven of them were aged 26–35, and three of them were between 36 and 45 years. Six participants were Asian (China, India, Indonesia, and Hong Kong), three African (Nigeria, and Ghana), and one European (British). Participants’ travel frequency ranges from once a year to eight times per year. Most MMEs reported were from leisure travel with a few from study and missionary trips. Two-thirds of the MMEs reported were less than five years. The whole study was conducted online with the help of the conference and instant messaging software for support and interview. Each participant earned a compensation coupon equivalent to US $10 after finishing all the steps. The research team contacted participants, explained to them the procedures involved, and checked with the participants their MMEs and materials (e.g., photos and information). Then the participants were invited to identify their character strengths using the online VIA-IS survey. The research team created a user account with the character strengths profile of each participant. Participants were invited to create at least five MMEs stories from their past journeys within two weeks. Fig. 3 shows an example of MME entry by Ron. After the journaling task, the participants were asked to reflect on their experiences by browsing the dashboard (Fig. 2(c)) and insight cards (Fig. 2(d)–(f)). They were prompted to rate their behavioural intentions related to any future trip plans (Fig. 2(g)). The data collection was finalised with a short online semi-structured interview (10–15 min in length) discussing the rationale for their responses and their experiences of the interventions. The study resulted in four sets of data for the analysis: 1) 10 character strength profiles; 2) 51 strengths-based MME entries (one participant created six entries); 3) 471 responses on behavioural intentions about future trips; and 4) the participants’ feedback on their choices made. Interviews were conducted in English and transcribed verbatim for thematic analysis. The process involved familiarisation with the data, coding, generating themes, reviewing and defining themes, as outlined by Braun and Clarke (2006). Participants’ quotes were used to illustrate the themes identified. Fig. 1. The strengths-based journaling platform: user profile and story creation (Wan et al., 2021). C.K.B. Wan et al.

Journal of Hospitality and Tourism Management 57 (2023) 61–71 65 4. Data analysis and findings 4.1. Strengths involved in the production of MMEs The first research question aimed to understand how do tourists draw upon their character strengths in their MMEs (RQ1). Three types of strengths were identified according to their significance to one’s life: the signature strengths (i.e., the top five strengths of one’s VIA profile), the moderate (middle) strengths (i.e., the sixth to nineteenth strengths), and the lesser strengths (i.e., the bottom five strengths). Since participants can register up to two character strengths for each entry, the strengths involved in the production of MMEs were categorised into nine groups (Table 3). The first four groups (1–4) consisted of 17 entries (33%) which registered at least one signature strength of participants. The next three groups (5–7) accounted for 30 entries (59%) which affiliated to moderate strengths. The last two groups (8–9) composed of four entries (8%) that involved the lesser strengths of participants. 4.2. Prominent strengths used and values gained Regarding the frequency of strengths used, the top five strengths used are the appreciation of beauty and excellence (23.5%), curiosity (21.6%), gratitude (17.6%), teamwork (17.6%), and perseverance (13.7%) (see Table 4). The distribution is affected by the type of tourism activities documented in the entries. For instance, the appreciation of beauty and excellence is related to contact with nature, discovering cityscape, and exploring historical sites. The strength of curiosity was mostly associated with the first-time encounter with local culture, immersing themselves in the local history, interacting with locals. The strength of gratitude, on the other hand, was of thankful feelings towards special encounters with people, service, food, and nature. The result of these encounters gave participants a sense of privilege and comfort. The strength of teamwork was mostly related to facing and tackling challenges with others for accomplishing more challenging and purposeful missions. At times, it also related to unexpected incidents that happened over their journeys. Lastly, the strength of perseverance entailed sports and outdoor activities under adverse conditions. The participants could associate a maximum of two values per each entry. The values indicated why the experience was perceived as memorable and meaningful to the participants. A total of 86 values were attributed to 51 MME entries (Table 5). Around one-fourth of stories were affiliated with knowledge (23.5%). Friendship and personal growth both shared 21.6%. Harmony accounted for 19.6%, and meaning and purpose in life accounted for 15.7%, followed by excitement (13.7%) and better world and society (11.8%). The data suggested that gaining new knowledge is highly appreciated by the tourist group. Gaining friendships and personal growth are very important aspects of MMEs. Personal development also seems to be a key aspect of MMEs and it may offer new opportunities for people to formulate growth goals for their future journeys. Interestingly, excitement, which is often considered as an important factor in tourism experiences only received 13.7%. The lower percentage may be due to the fact that the participants were asked to report on their memorable as well as meaningful travel experiences, instead of solely being asked to report on their memorable Fig. 2. The strengths-based journaling platform: story browser and reflection. C.K.B. Wan et al.

Journal of Hospitality and Tourism Management 57 (2023) 61–71 66 experiences, as in prior studies (e.g., Chandralal & Valenzuela, 2015; Kim et al., 2012). 4.2.1. Reminiscing and behavioural intentions The strengths-based journal can make the implicit psychological dimension (i.e., character strengths) of MMEs explicit via a structured narrative. As a result, participants would be aware of their character strengths gained from their MMEs. Therefore, RQ2 examined participants’ willingness to do the following: 1) pursue the values gained, 2) develop the strengths used, 3) travel with the same person, 4) do the same activities, and 5) visit the same place in their future journeys after creating MMEs entries. Overall the above 5 points addressed the broader goals of RQ 2 which were to understand how do tourists take their MMEs into consideration when planning future trips. On 51 stories created, the platform generated 131 questions on the values gained, 107 questions on the character strengths used, and 131 questions on the relationship created, 51 queries on the activity taken and place visited. Fig. 4 shows the means score of participants’ responses while the error bar displays the standard deviation of the questions answered. The shorter error bar indicates consensus among participants Fig. 3. A sample of MME entry provided by Ron (P10) (Wan et al., 2021). Table 2 List of participants. Code Pseudonym Gender Age Nationality Profession Frequency of travel (per year) P1 Susan F 26–35 Hong Kong (China) Medical Scientist 3–5 P2 Cherry F 26–35 Hong Kong (China) NA 6–8 P3 Jane F 46–55 Canadian Programme Officer 1–2 P4 Allan M 20–25 Ghanian NA 1–2 P5 Florian M 26–35 Ghanian Postgraduate Student 1–2 P6 Pauline F 26–35 Indonesia Postgraduate Student 3–5 P7 Yolanda F 26–35 Hong Kong (China) Designer 8+ P8 David M 36–45 British Educator 6–8 P9 Sylvia F 26–35 Indian Researcher 3–5 P10 Ron M 36–45 Nigerian Computing 3–5 C.K.B. Wan et al.

Journal of Hospitality and Tourism Management 57 (2023) 61–71 67 on the questions. The result suggested that participants were more willing to pursue what they had found valuable to them. Many of them were willing to develop the strengths they had used in their past journeys but were less determined to take a trip again with the travel mates they had traveled with in the past. Place re-visitation was the least favorable option among the five experiential attributes. Thematic analysis was then conducted to capture the participants’ perceptions of the interventions (Braun & Clarke, 2006). The result showed three benefits from using the platform: 1) gaining new insights from MMEs, 2) becoming more aware of character strengths, and 3) supporting decision making for future journeys. 4.2.1.1. Gaining new insights from MMEs. Participants found the reflection section (Fig. 2 (c)) helped them gain insights into their innate needs and capabilities. For instance, Jane (P3) mentioned that the reflection section helped her to plan her trip with the focus on developing her strength of leadership, which she had never thought of before. Allan (P4) found the reflection section illuminating. She revealed that the trips she made revealed her sense of personal growth. “These experiences really taught me a whole lot of lessons in life and made me become who I am today” noted Allan. 4.2.1.2. Becoming more aware of character strengths. Participants (P1, P3, P4, P5, P7, P8, P9) gained greater willingness to cultivate their character strengths and pursue meaningful goals. This broadened their minds and allowed them to have more choices when planning for their future journeys: “I will now consider my character strengths and values into my future trips” (Allan, P4). David (P8) was excited to know that the experiences were connected to his character strengths: “I now understand why I treasure these experiences so much!“. 4.2.1.3. Supporting decision making for future journeys. Reminiscing about MMEs with character strengths in mind triggers tourists to embark on future trips with a clearer set of goals and intentions that are congruent with their character strengths. Cherry (P2) stated that she would prioritse love of learning and appreciation of beauty and excellence (her signature strengths) instead of seeking autonomy when planning future trips. In contrast, Susan (P1) valued the sense of autonomy more than seeking a better world and society (one of the values she realised in her past journey) and she will utilise this sense of autonomy in the future. Ron (P10) claimed that he had always been planning trips with his personality traits in mind. 5. Discussion and implications 5.1. Discussion There is little doubt that tourism provides opportunities for tourists to create MMEs. This study facilitated tourists to gain a deeper awareness of their MMEs by connecting them with their character strengths through reminiscence, addressing both RQs 1 and 2. The reminiscence process involved using a strengths-based journaling platform to create MME stories that connected explicit tourism activities (people, places, activities, peak moments) with implicit psychological dimensions (emotions, character strengths, values). Reflecting on these experiences allowed the participants to identify behavioural and psychological patterns and express their future travel intentions. The results revealed that Table 3 Categories of character strengths used. No Strength Category Frequencya (N = 51) Percentb 1 Two signature strengths 2 3.9 2 One signature strength 7 13.7 3 One signature strength and one moderate strength 6 11.8 4 One signature strength and one lesser strength 2 3.9 5 Two moderate strengths 7 13.7 6 One moderate strength 15 29.4 7 One moderate strength and one lesser strength 8 15.7 8 One lesser strength 3 5.9 9 Two lesser strengths 1 2 Total 51 100 a The participants could attribute a maximum of two character strengths to each MME entry. b Percentage is calculated based on the total number of entries (N = 51). Table 4 Frequency of strengths used across stories. Strengths P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 Freq.a Pct.b Appreciation of Beauty & Excellence 2 1* 1 1 1 2 – 2* 1* 1* 12 23.5 Curiosity 1 1 2 1 2 1* * * 1* 2 11 21.6 Gratitude 2 1* 3 1* – 1 – – 1 * 9 17.6 Teamwork 1 1* * 1 1 * 1 2 1 1* 9 17.6 Perseverance – – 1* 1 * – 1 – 1* 3 7 13.7 Love of learning 1* 1* – 1* 1 – * – 2 – 6 11.8 Love 1 – – – 1 – 1 1 1 5 9.8 Self-Regulation * – – – – – 2 – – 1 3 5.9 Bravery – – 1 – – – – – – 1 2 3.9 Social intelligence – – – – – * 1 – 1 – 2 3.9 Spirituality – * 1 * * * 1* * – – 2 3.9 Creativity – – – – 1 – – – – – 1 2 Forgiveness – 1 – – – – – – – – 1 2 Honesty – – * 1* * – – – * – 1 2 Humility * 1 – – – – – – – – 1 2 Kindness 1 – – * – * – * * * 1 2 Leadership – – 1* – – – – – – – 1 2 Perspective 1 – – – * – – – – – 1 2 Prudence * 1 – – – – * – – – 1 2 Zest – – – – – – – – 1 – 1 2 Fairness – – * – – – – – – * 0 0 Hope – – – – – – * – – – 0 0 Humor – – – – – – – * – – 0 0 Judgement * – – – * – – – – – 0 0 a Each MME entry could be associated with a maximum of two character strengths. b Percentage is calculated based on the total number of entries (N = 51) [*] Signature character strength. C.K.B. Wan et al.