Exploring How Tourist Engagement Affects Destination Loyalty: The Intermediary Role of Value and Satisfaction

: A conceptual network between and among tourist engagement, perceived value, tourist satisfaction, and destination loyalty was constructed and validated based on 401 valid samples collected in Phoenix Ancient Town, a famous destination in western Hunan, China. The research results conﬁrm the logic chain in marketing psychology of “engagement-value-evaluation/behavior” in the context of a tourism destination. Speciﬁcally, tourist engagement has direct positive effects on destination loyalty, but not on tourist satisfaction. In addition, tourist engagement exerts indirect positive effects on destination loyalty by producing functional value and emotional value for tourists which, in turn, promotes tourist satisfaction towards a destination. However, the indirect inﬂuences of tourist engagement on tourist satisfaction and destination loyalty by way of creating tourists’ social value are not signiﬁcant. This research validates the role of tourist engagement in the formation of destination loyalty from a tourist value perspective and thereby offers a new theoretical clue for tourism marketing.


Introduction
In the last decade, consumers have been increasingly communicating with both marketers and other consumers through mobile internet such as social media, online travel service providers, and mobile search websites, etc. [1]. A new trend of online and offline interaction has emerged in the tourism and hospitality industry. Various industries and academic institutions emphasize that throughout the process of multi-media and multichannel service contact, customer involvement is important for companies to obtain new competitive advantages [2,3].
Theoretically, as the key construct to highlight the interactive, co-creative experiences between customers and service providers, the term of "customer engagement (CE)" has been widely used in discussions addressing service-dominant logic [4] and the expanded domain of relationship marketing [5]. Academic research has gradually recognized the important role of CE in predicting customer emotions and behaviors, such as customer trust [6,7], customer advocacy [8], customer value co-creation [9], customer satisfaction [10,11], and customer loyalty [12][13][14].
The lively discussion about the impact of CE has been expanding to more consumption areas including leisure and tourism. Traveling is a highly engaging, enjoyable, emotional, and visual process, which can arouse the tourist's discussions online. In addition, tourists rely on the knowledge and creativity of other tourists in their travel experiences, making tourism context unique in cultivating CE [15]. Accordingly, researchers have attempted to develop a tourism and hospitality-based CE scale [16]. Based on the scale, So et al. [17] examined the effects of CE on customers' loyalty to hotel and airline brands. Harrigan et al. [18] extended the scale of So et al. (2014) in the context of social media and examined its impact on tourists' behavioral intention and loyalty. Yen et al. [19] studied the influence of CE on customer co-creation behaviors in restaurants. Godinho et al. [20] developed a scale to examine CE's relationship with the customer's involvement, emotional state, experiences, and brand advocacy in hospitality industry.
Previous studies mainly focus on CE within companies or online communities, which makes destination-based CE lacking exploration. A destination involves multiple objects, including but not limited to service providers, tourist attractions, tourists, travel companions, and local residents [21], all of which can initiate an engagement. In addition, visitors are no longer satisfied with being passive, but are willing to contact with various objects in an experiential and interactive online consumption environment [5]. Tourist engagement at a destination has been a hot topic in academic research [15,[21][22][23]. To reveal the impact mechanism of tourist engagement (TE), past studies have explored the relationship between TE and tourists' post-visit evaluations and behaviors (e.g., loyalty, memorable tourism experience) through intermediary variables of brand/customer trust [17,24] and cultural contact [25]. However, the perspective of customer value has been overlooked. The creation of customer value and the subsequent satisfaction and loyalty level are critical in marketing [26]. In addition, CE plays an important role in value co-creation [9,24], making the inter-correlations among TE, perceived value, tourist satisfaction (TS), and destination loyalty (DL) an interesting topic to be further studied.
To bridge the research gaps, this study employs a TE scale developed by Wang, Xu and Wang [23] within a Chinese destination context and empirically examines the driving force of TE towards TS and DL through perceived value and its specific dimensions. The rest of the paper is organized as follows: Section 2 expands the theoretical background and proposes hypotheses based on previous literature; Section 3 specifies the empirical model and data; Section 4 tests the hypotheses using data obtained from Phoenix Ancient Town and presents the analysis results; and Section 5 concludes this research and provides implications, limitations, and future research directions.

Tourist Engagement: Conceptualization and Dimensionality
Marketing consultants and scholars have explored CE within diverse contexts (e.g., online media, service industries, consumer brands), and interpreted CE from seven main perspectives (i.e., connection, interaction, behavior, loyalty, experience, commitment, process) with three common special attributes. First, CE is a non-transactional relationship, which is different from traditional relationships (e.g., participation, involvement) from a narrower perspective [27]. Second, CE is based on interactive experiences and value co-creation [4,5]. Third, CE is induced by consumers' intrinsic motivations, which indicates active customer relationships with focal objects (e.g., brands) [28,29].
CE is treated as a single or a multidimensional construct. Some studies used unidimensional scales to measure CE, mainly focusing on behavioral aspects [30][31][32]. However, most researchers argue that CE should be constructed from both psychological (i.e., cognitive and emotional) and behavioral aspects [33]. Vivek [34] developed a scale of CE consisting of enthusiasm, conscious participation, and social interaction. Hollebeek et al. [35] developed a social media engagement scale, including cognitive processing, affection, and activation.
During recent decades, a widespread effort has been made to develop engagement scales in the tourism and hospitality context. So, King, and Sparks [16] developed and validated a CE scale with airline and hotel brands by referring to the dimensions of work engagement and CE, which includes five dimensions of enthusiasm, attention, absorption, interaction, and identification. Harrigan, Evers, Miles, and Daly [18] developed an alternative three-factor (i.e., identifification, absorption, interaction), 11-item CE scale in a social media context based on So, King and Sparks [16]. These CE scales are useful but cannot be generalized to the destination context, as the scales do not accurately reflect the behavior, emotions, and experiences of tourists in a destination [21]. To address this issue, Huang and Choi [21] developed a four-dimensional TE scale in the context of cruise travel, including social interaction, interaction with employees, relatedness (i.e., the connectedness with travel companions) and activity-related engagement.
Wang, Xu and Wang [23] used the perspective of connection to explore the definition and measurement of TE in a Chinese destination, and found out that TE has five dimensions: enthusiasm, attention, absorption, social interaction, and identification. The connection perspective is not only compatible with the above three common special attributes, but also inherits the corn of employee engagement (i.e., the full and effective connection between employees and work) [36], which makes it the most widely applied definition in the field of marketing, especially in the context of leisure, tourism, and hospitality [5,16,21,23].

Research Model and Hypotheses Development
Our conceptual model is originally linked to the regulatory engagement theory in psychology. The theory states that engagement is a state of involvement and full absorption [37], which argues that value not only comes from the hedonic experience, but also from the strength of engagement through the experience of motivational force, i.e., the pull or push force of the value target [38]. Further, individuals' evaluations and behaviors (e.g., task performance) are affected by engagement, which indicates that the more engaged an induvial is, the more positively he/she pursues the target [28]. Student engagement is closely related to study performance [39], and brand engagement improves organizational trust, commitment, and service brand loyalty [7]. Engagement expands to the marketing field grounding on premises of service-dominant logic [4,40] and expanded domain of relationship marketing [5]. Through co-created, non-transactional, and interactive experience generated in the engaging process in the networked service relationships, customers obtain various experiences and multiple values [4,5,41]. According to the theories mentioned above, a value creation logic of "engagement-value-evaluation/behavior" is hypothesized in a destination context. Specific research hypotheses of the relationship between TE and the outcome variables are proposed based on previous literatures.

Tourist Engagement and Perceived Value
Perceived value is customer's perceived preferences and evaluations of product attributes and functions, which help or hinder the achievement of customers' goals under specific scenarios [42]. Perceived value of a customer is composed of emotional value, functional value, and social value [43]. The driving factors of value include experience/service quality [44], motivation, and involvement [45]. Marketing literature highlights the role of CE in co-creating value [4]. Hollebeek [46] concluded that brand engagement is positively correlated with customer value by using an in-depth interview and focus group interviews.
Research conducted on the direct relationship between engagement and the dimensions of perceived value is currently lacking. However, relevant research indicated that CE has a positive impact on the perceived quality and the price of tourism-related services [17,34]. In addition, Hollebeek [46] showed that hedonic brand engagement enhanced brand users' symbolic value (including social value), experiential value, self-image, and self-value. Furthermore, engagement, based on its relational foundations of interactive experience and value co-creation [4], will probably affect tourists' functional value, emotional value, and social value positively through their co-creating experiences with the destination [47]. Empirical studies also demonstrated that the active interactions among tourists positively affects their functional value, emotional value, and social value [48]. Based on the above analyses, the following hypotheses are proposed.
Hypothesis 1a (H1a). TE has a significant positive impact on functional value. Hypothesis 1b (H1b). TE has a significant positive impact on emotional value.
Hypothesis 1c (H1c). TE has a significant positive impact on social value.

Tourist Engagement and Tourist Satisfaction
Since the 1950s, customer satisfaction has gradually become a core indicator for evaluating the quality of products. Oliver and Linda [49] believed that satisfaction is an emotional state that occurs when customers' expectations are consistent with or exceeds their actual consumption experiences. Patterson and Yu [10] contended that CE will promote customer satisfaction. In the tourism industry, CE in tourism brands has been indicated to positively influence hospitality brand evaluation (including satisfaction) [17]. As a kind of identity with a destination, perceived destination eco-friendliness reputation was proved to be an important determinant of TS [50]. According to the regulatory engagement theory, the higher the engagement level, the greater the driving force for the focus object, thereby boosting more positive experiences [37]. This indicates that highly engaged tourists are more likely to obtain a pleasant experience at destinations, which promotes TS. Empirical studies have uncovered that social interactions have a significant positive impact on TS [51]. Accordingly, this study proposes: Hypothesis 1d (H1d). TE has a significant positive impact on TS.

Tourist Engagement and Destination Loyalty
Cultivating loyal customers is the core of relationship marketing. Marketing researchers pointed out that CE directly enhances customer loyalty [13,24,52], and more engaged consumers are more motivated to recommend products/brands [14,53]. Engagement can enhance an individual's commitment or positive attitude towards products, brands, or enterprises eventually improving brand loyalty [5,7] or DL [54]. Identity, one of the important dimensions of CE, presents a positive impact on loyal behaviors such as positive word-of-mouth [55,56] and intent to visit [57]. Interaction with travel companions will evoke tourists' pleasant emotions and thus promote their storytelling intentions [57]. Accordingly, this study proposes: Hypothesis 1e (H1e). TE has a significant positive impact on DL.

Perceived Value, Tourist Satisfaction and Destination Loyalty
Marketing research has basically confirmed the logic chain of "value-satisfactionloyalty" [42,58]. Despite customers' perceptual objects varying, the positive impacts of perceived value on TS are generally verified [59,60]. Woo et al. [61] concluded that residents' perceived value of tourism development positively affects satisfaction. Kim and Park [62] validated the positive effects of overall perceived value, (TS on DL) on community-based eco-tourism. Williams and Soutar [63] argued that functional value and emotional value have significant positive impacts on TS. Gallarza and Saura [64] concluded that social value significantly positively affects TS. Consequently, the following hypotheses are proposed:

Hypothesis 2a (H2a). Functional value has a significant positive impact on TS.
Hypothesis 2b (H2b). Emotional value has a significant positive impact on TS.
Hypothesis 2c (H2c). Social value has a significant positive impact on TS.
Sweeney and Soutar [43] found that functional value, emotional value, and social value exert a significant positive impact on consumers' willingness to purchase and recommend. Several studies revealed that there is a significant positive relationship between perceived value, emotional value, and tourists' behavioral intentions [63,65]. Chi et al. [66] confirmed the positive effects of tourist-perceived quality on behavioral intentions in rural tourism. Accordingly, this study proposes: Hypothesis 3a (H3a). Functional value has a significant positive impact on DL.
Hypothesis 3b (H3b). Emotional value has a significant positive impact on DL.
Hypothesis 3c (H3c). Social value has a significant positive impact on DL.
Consumer satisfaction is widely recognized as an antecedent of loyalty in the marketing arena. It is well documented in tourism studies that TS significantly affects DL [22,67,68]. The following hypothesis is therefore proposed.

Hypothesis 4 (H4). TS has a significant positive impact on DL.
Based on the above-mentioned research hypotheses, a conceptual model is constructed ( Figure 1).

Measurement Scale
In our study, TE is defined as the cognitive, emotional, and beyond-purchase behavioral connection between a tourist and a destination, which is generated by tourists in the interactive travel experiences related to destination attractions, services, and activities [21,23]. This experience includes interaction through destination marketing channels (e.g., destination online communities) or interaction with people at the destination (e.g., destination service personnel). The strength of the connection is the level of engagement. The TE scale with five dimensions and 25 items developed by Wang, Xu, and Wang [23] is used in this study. The operational definitions of the five dimensions are as follows [21,23]. Attention means that tourists are consciously concerned about the destination and participate in destination tourism services and activities. Enthusiasm expresses the interest and excitement of tourists towards the destination and its services and activities. Absorption embodies the state of concentration of tourists with the destination and its related personnel and media. Identification means the commitment and emotional connection between the tourist and the destination. Social interaction refers to tourists' psychological feelings when they participate in or discuss destination related travel activities online or offline.
The above TE definitions and scale were selected for the following reasons: (1) engagement is conceptualized and measured from connection perspective, which reflects the key attributes (comprising cognitive, emotional, and behavioral components) of CE. Attention and concentration belong to the cognitive component, enthusiasm and identification belong to the emotional component, and social interaction is the behavioral component; (2) this scale was developed based on the prior researches of So, King, and Sparks [16], and Huang and Choi [21] in the arena of hospitality and cruise travel; (3) the concept and scale have been tested in Chinese destinations, which verified the reliability and validity of the scale in Chinese destination context.
Perceived value consists of three sub-dimensions (i.e., functional value, emotional value, and social value) and 14 items obtained from Sweeney and Soutar [43] and Williams and Soutar [63]. A four-item TS scale from Oliver [69] is adopted to measure tourists' overall satisfaction towards a destination. DL is measured using five items from Hutchinson et al. [70]. In order to ensure that the analysis sample has a relevant online interactive experience, the questionnaire designs a sample screening question, that is, whether you have shared and exchanged information related to Phoenix Ancient Town through the mobile internet (e.g., social media, online travel communities). The answer options are divided into yes or no, and the valid samples are limited to responders who choose "Yes".

Research Area
Phoenix Ancient Town in Hunan Province, China was selected to conduct an empirical examination of the conceptual model. Phoenix Ancient Town is China's national historical and cultural city, a national 4A-level scenic spot, and is praised by the New Zealand writer Rewi Alley as one of the two most beautiful towns in China. The town is rich in tourist attractions and features China's Miao minority culture. In 2019, Phoenix Town was visited by 209.39 million tourists with a tourism income of CNY 20.01 billion [71]. After more than 30 years' development, the town has become a well-known destination for travel in China, with diverse experience activities such as boating on the Tuojiang River, rock jumping, bonfire parties, entertainment in bars, and shopping in ancient streets, etc. Due to the boom of the tourism industry, Phoenix Ancient Town's tourism products have been widely promoted by offline and online travel agencies, online portals, and search engines, providing tourists with ample interaction and co-creation experiences.

Data Collection and Sample Profile
In July 2019, a self-report questionnaire was used to collect data. A total of 400 questionnaires were distributed to individual tourists by the authors, and 220 questionnaires were distributed to group tourists by tour guides. To ensure the consistency of the samples, the respondents were restricted to tourists who had just finished visiting Phoenix Ancient Town. 574 questionnaires were returned, of which 401 questionnaires were identified as valid after discarding incomplete and unqualified questionnaires. SPSS 21.0 was employed to conduct statistical analysis. First, the two-sample Kolmogorov-Smirnov (K-S) test was used to examine whether the samples collected from both sources conformed to the same general distribution. The two-sample K-S test is an effective method for comparing the distribution, as it is sensitive to the shape and position parameters of the distribution function. The test shows that the null hypothesis is rejected at the 5% significance level, indicating that there is no significant difference between the distribution of the two samples, i.e., they belong to the same population. Second, Harman's single factor method was adopted to test the variables' common method bias. Specifically, the dimensionality reduction of all variables was conducted using unrotated exploratory factor analysis. If only one factor is extracted, it suggests that the variance contribution of the factor is abnormally large, and the common method bias is serious. Eight factors were extracted in our study, with a 43.7% variance contribution rate of the first factor, and a cumulative variance contribution rate of 73.8%, demonstrating that there was no serious common method bias of the measurement scale.
Descriptive statistics analysis shows that 50.6% of the respondents were male and 41.4% were female, with 32 missing values. Most of the respondents (45.9%) were between 15 and 24 years old, with 1.5% younger than 15 years old, 25.4% aged from 25 and 34 years old, 13.5% between 35 and 44 years old, 6.7% between 45 and 59 years old, and 2.5% older than 60 years old. The majority of the respondents had received at least high school or undergraduate school education, accounting for 20.7% and 56.9%, respectively. The proportion of respondents who had received a primary school education or below, middle school, postgraduate, and above were 1.0%, 2.5%, 4.7%, 14.2%, separately. In terms of average pre-tax monthly income, 24.7% of the respondents had no source of income, 2.5% earned less than RMB 1000 per month, 16.2% earned between RMB 1001 and RMB 5000 per month, 11.7% earned between RMB 5001 and RMB 8000 per month, 6.5% earned between RMB 8001 and RMB 10000 per month, and only 1.7% of the respondents had a monthly salary over RMB 10001. The occupation of the respondents was distributed across 21 industries, many of whom were students (29.2%). In addition, 67.6% of the respondents were first-time tourists and 26.7% of them were repeat tourists.

Results of Hypotheses Testing
Confirmatory Factor Analysis (CFA) was used to test the model fit of TE, perceived value, and the measurement model (Table 1). Reliability is evaluated by α coefficient and internal consistency. Convergent validity is manifested by the standardized factor loadings of variables, significance level (P), and Average Variance Extracted (AVE). If the standardized loadings are between 0.5 and 0.95, the p-value is smaller than 0.05, and the AVE value is greater than 0.5, it indicates a good convergent validity. If the correlation coefficients of the latent variables are smaller than the square root of AVE, the discriminant validity between the latent variables is good. A second-order CFA model of TE was constructed following So, King, and Sparks [16]. As presented in Table 1, all indices meet the standard, with the exception of GFI values, which are slightly lower than the cutoff value. Bentler [72] pointed out that when CFI is greater than 0.9, the critical value of GFI could be lowered to 0.85. Accordingly, the results show a good fit of the model with the data. According to Table 2, the standardized factor loading of each variable is between 0.589 and 0.946, the p-value is less than 0.001 (two-tailed), and the AVE values of four dimensions are greater than 0.5. The AVE value of "enthusiasm" is slightly smaller than 0.5, which is acceptable according to Welker. [73]. The above results indicate a good convergent validity of the TE model. Table 3 shows the square roots of AVE values of the five dimensions ranging from 0.670 to 0.784. Their correlation coefficients range from 0.534 to 0.733, which are all smaller than the square roots of AVE, indicating a good discriminant validity of the latent variables. The CFA results in Table 2 show that the TE scale has a good reliability and validity.
Following Sweeney and Soutar [43], a first-order CFA model of perceived value was adopted. Table 1 shows that the perceived value model fits well with the data. According to Table 4, the standardized factor loadings of variables range from 0.675 to 0.935. T-values range from 13.605 to 20.326, and p-values are smaller than 0.001 (two-tailed). AVE values of emotional value, social value, and functional value are 0.711, 0.761 and 0.556, respectively, which are all greater than 0.5, indicating a good convergent validity. In addition, the α coefficients of the three dimensions are 0.907, 0.933 and 0.889, respectively, and the composite reliability is 0.908, 0.927 and 0.882, respectively, demonstrating a good reliability of the perceived value measurement model. In terms of discriminant validity, the correlation coefficients of the three latent variables are 0.663, 0.549 and 0.478, respectively, which are all smaller than the square root of the AVE values, indicating a good discriminant validity between and among the latent variables. The above results validate the first-order threedimensional model of perceived value. The analysis results of the overall measurement model also show a good model fit with the data (Table 1). In Table 5, the standardized factor loading of each item is between 0.622 and 0.931, exceeding the critical value of 0.5. T-values range from 12.164 to 28.183 and are significant at 1‰ significance level (two-tailed). The AVE values range from 0.547 to 0.762, which exceed the critical value of 0.5, indicating a good convergent validity. Furthermore, the α coefficients and the composite reliability of the latent variables are all greater than the recommended level of 0.7, indicating a good reliability of the variables.  Table 6 shows that the correlation coefficients between the six latent variables are all smaller than the square root of the corresponding AVE values, indicating a good discriminant validity between and among the latent variables. Combining the results in Tables 5 and 6, the overall measurement model has a good reliability and validity.   As depicted in Figure 2, the path analysis results of the conceptual model show that the standardized regression coefficient (β) of TE on functional value is 0.763, which is significant at 1‰ significance level (two-tailed), confirming H1a. In addition, TE is found to exert positive effects on emotional value and social value (β = 0.732, p < 0.001; β = 0.726, p < 0.001), verifying H1b and H1c. TE is proved to significantly and positively affect DL (β = 0.261, p = 0.011), which confirms H1e. However, H1d is failed to be confirmed (β = 0.149, p = 0.127). Functional value and emotional value show positive influence on TS, confirming H2a and H2b. Functional value, emotional value, and TS present positive impacts on DL, verifying H3a, H3b, and H4. However, the positive effects of social value on TS and DL are not significant. Therefore, the data failed to confirm H2c and H3c. Figure 2. Hypothesis testing results. Note: H represents hypothesis; ***, ** denote hypotheses accepted at 1‰ and 1% significance level, respectively (two-tailed). The dashed lines denote a failure to verify the hypotheses.

Discussions and Implications
A conceptual model based on marketing psychology theories (e.g., regulatory engagement theory) and multi-media interactions was constructed in a destination context by incorporating TE, perceived value, TS, and DL. It is discovered that tourists' engagement psychology and behaviors are positively associated with tourists' post-visit evaluations and behaviors via creating tourist value. The specific empirical findings are as follows.
First, TE is a significant predictor of perceived value, which is similar to the research results of Vivek [34], who carried out research based on service and consumption activities. Specifically, TE significantly affects functional value, emotional value, and social value, with standardized coefficients of 0.763, 0.732 and 0.726, respectively. This indicates that one standard deviation change of TE leads to 0.763, 0.732 and 0.726 standard deviation changes of functional value, emotional value, and social value, separately. This result suggests that the TE process contains tourists' motivation-driven and interactive experiences with a destination online or offline. This confirms the regulatory engagement theory [37,38] in a destination context, i.e., TE with a destination is another source of tourists' perceived value in addition to the hedonic experiences.
Second, TE positively influences TS through functional value and emotional value. The total indirect standardized coefficient of TE on TS is 0.583, which demonstrates that TE induces functional value and emotional value, thereby improving TS. The role of TE in cocreating parts of values and satisfactory experiences of tourists visiting a certain destination has been validated. However, the significant mediating role of social value between TE and TS is absent, which is different from the finding of Gallarza and Saura [64], but is consistent with the research of Li [74] on ancient villages. The possible explanation for this result may be that Phoenix Ancient Town is a popular but low-priced sightseeing-dominated destination, which inhibits tourists' perceptions of social value during the journey (i.e., improving social self-identity and making a good impression on others, etc.).
Third, TE positively affects DL both directly and indirectly. The direct effects of TE on DL is 0.261. Previous studies indicated the significant positive impact of CE on brand loyalty by using a CE scale [24], or a hotel and airline brand engagement scale [17]. This study has further confirmed the positive effects of TE on DL by using a TE scale. As pointed out by Prebensen and Foss [75], engaged tourists are more likely to have special and unforgettable experiences, which is positively correlated with their post-visit behavioral intentions. Additionally, TE positively affects DL through functional value, emotional value, and TS indirectly, with a total indirect standardized coefficient of 0.525. Inconsistent with the results in a retail context [43], but similar to the empirical findings on adventure tourism [63] and ancient villages [74], the direct effects of social value on DL is not verified in the study, indicating that the relationship between social value and loyalty may vary across different research contexts.
Fourth, since the direct effects of TE on TS is not significant, the mediating role of TS between TE and DL is not confirmed. This result is inconsistent with Patterson and Yu [10] and Rasoolimanesh, Md Noor, Schuberth, and Jaafar [22], who maintained that customer engagement had positive effects on customer satisfaction. It suggests that the effects of engagement on satisfaction vary across both different industries and destination types. An in-depth investigation into this relationship and influence mechanism may offer fine-grained explanation for the heterogeneity.
The above findings provide theoretical and practical implications for destination marketing strategies based on engagement theories. Theoretically, although satisfaction is a necessary step in the formation of loyalty, as loyalty begins to be established through other mechanisms, satisfaction becomes less important [76]. The current research examined these other mechanisms, with special attention to the impact of TE on DL. A theoretical network between TE, perceived value, TS, and DL was validated, which can be summarized as the logic chain of "engagement-value-evaluation/behavior". This is in line with the engagement effect in the fields of psychology and behavioral research [7,39]. TE's impact mechanism on DL promotes the empirical application of regulatory participation theory, service-dominant logic, and expanded relationship marketing in destination scenarios, with a view to establishing the relationship between TE and other traditional destination marketing variables.
Practically, as TE positively influences DL directly and indirectly via functional value, emotional value, and TS, destinations should take measures to create functional value and emotional value for tourists by improving TE, thereby ultimately improving TS and DL. To obtain tourists' identification, attention, enthusiasm, social interaction, and absorption, communicating channels (e.g., online communities, official WeChat account or applets, Weibo accounts and online games, etc.) between tourists and destinations beyond purchase should be designed and actively operated to facilitate tourists' real-time interactions with destinations. The destination should establish a unified, interactive, credible, and interesting mobile network medium to improve the TE. Second, both regular and updated experience activities, such as traditional tourist events and festivals and VR-based activities can be adopted to trigger tourists' social interaction, attention, enthusiasm, absorption, and identification towards a destination. Last, as functional value is an important mediating indicator between and among TE, TS, and DL, destinations should constantly optimize their public and private service facilities and quality.

Limitations and Future Research Directions
This study contributes to the existing tourism marketing literature by constructing and empirically testing a conceptual model between TE, perceived value, TS, and DL. However, it has the following limitations. First, questionnaires were distributed to individual and group tourists separately, but the TE mechanisms of the two tourist groups were not compared. Second, data collection was implemented only in a Chinese destination. The country-specific and culture-specific samples lead to the limitation in generalizing research findings to other countries or cultures.
Several future research directions are suggested by the current study. First, this research explores the consequences of TE. In fact, exploring the antecedents (e.g., motivational or relational drivers) of TE has been a topical issue and can be an exciting direction for follow-up studies. Second, TE for first-time and repeat tourists may be different. For example, some variables are antecedents of TE for first-time tourists, but may be consequences for repeat tourists. Accordingly, it is necessary to conduct comparative research between the two groups of tourists, thereby revealing different influencing paths of TE. Third, a destination is a collection of diverse servicescapes. Therefore, it will be interesting to conduct more focused investigations of TE in specific servicescapes, such as tourist scenic areas, theme parks, and online travel communities, etc. Finally, Phoenix Ancient Town is a comprehenve scenic area comprising both heritage and non-heritage scenic spots, which makes fine-grained investigations into different types of tourists and scenic spots promising topics in the future.  Data Availability Statement: All the data in this study were generated by questionnaires, which are available upon request.

Conflicts of Interest:
The authors declare no conflict of interest.