How servicescape unleash customer engagement behaviors through place attachment: an investigation in a non-Western context

Abstract Little empirical research has delved into the impact of servicescape on customer attachment and customer engagement behaviors. This study investigates two main research questions in a non-Western country. First, this research asks how upscale hotels’ servicescapes influence customer engagement behaviors, and second, it investigates how place attachment mediates the relationship between servicescape and customer engagement behaviors. Questionnaires were distributed to 405 guests from upscale Egyptian hotels and resorts, and the responses were analyzed with structural equation modeling. The findings support our proposed model and prove the mediating role of place attachment. The servicescape, directly and indirectly, influences customer engagement behaviors, given that place attachment plays a partial mediating role. The findings provide considerable insight to help hotels unleash customer engagement by creating servicescapes that encourage customer attachment.


PUBLIC INTEREST STATEMENT
In recent years, the concept of customer engagement behaviors has become one of the common buzzwords in marketing and is viewed as a key source of competitive advantage for companies. In addition, the servicescape emerged as one of the key factors affecting customer behavior. This study examines the role of servicescape on customer engagement behaviors through place attachment in a non-Western context. The results show that servicescape positively affects different forms of engagement behaviors. The results also reveal the key role of place attachment. Interestingly, the place attachment is found to create a sense of loyalty and active participation of customers in cognitive, affective, and behavioral activities. The study provides useful insights for both researchers and practitioners on the role of servicescape and place attachment in eliciting customer engagement behaviors.

Introduction
The transition to the experience economy has shifted travelers' perceptions of hotels from simple places to sleep to venues for memorable and meaningful experiences. Hotel visitors become excited, full of life, and eager to discover new things (Carrington, 2016). With this transition, customers are anticipated to spend 140% more when they have a positive hotel experience (Dedeoglu et al., 2018). Consequently, today more than ever, hospitality firms must connect with customers by engaging them as brand marketers (Choi & Kandampully, 2019).
In the last two decades, engagement has been a crucial concept in diverse social sciences, including sociology, psychology, political sciences, organizational behavior, and management sciences (Brodie et al., 2011;Hollebeek, 2011). In turn, customer engagement, a customer's interactive brand experience, is rapidly gaining considerable scholarly attention in tourism and hospitality literature (Brodie et al., 2011;Islam et al., 2019;Kumar & Pansari, 2016;Rather et al., 2019). This reflects the dynamic nature of the customer engagement concept, which discourses the consumer's cognitive, emotional, and behavioral interactions during service delivery (Hollebeek, 2011;Rather et al., 2019). There is a rivalry between hospitality firms in the same niche and service context in today's market. Customer engagement is shaped by core services, physical surroundings, and social interaction (Tran et al., 2020). Thus, in fierce competition, tourism and hospitality services providers should realize the importance of developing distinctive servicescape identities to reinforce customer engagement and trust (Rather et al., 2019).
The servicescape concept has been widely recognized in the service sector since the influential study of Bitner (1992) regarding the classical notion of physical servicescape. Later, Tombs and McColl-Kennedy (2003) introduced the concept of social servicescape. Studies suggest that tangible and intangible surroundings are essential determinants of customer experience, especially for hospitality, where guests interact more with physical and social servicescape (Tubillejas-Andrés et al., 2020;Xu & Gursoy, 2020). Servicescape is considered essential for tourism service providers and building and maintaining servicescape represents a substantial financial investment for hotels and resorts (Peng et al., 2020). In response to this growing importance of servicescape in the service sector, Dedeoglu et al. (2018) argued that research is needed to examine the underlying mechanism of how servicescape influences customer behaviors.
Concerning servicescape models, Tubillejas-Andrés et al. (2020) argued that Mehrabian and Russell's (1974) Stimulus-Organism-Response (S-O-R) model is a prominent approach to research the servicescape in consumer behaviors. This study proposes place attachment as a mediator between the servicescape and customer engagement behaviors. Recently, Loureiro and Sarmento (2019) concluded that tourists who feel emotionally attached become more engaged, even if it is their first visit. Furthermore, Tsai et al. (2020) argued that hotels must provide comforts such as a home combined with hotel amenities to enhance visitors' place attachment and nostalgic emotion. Place attachment has received much attention since the seminal paper of Low and Altman presented in 1992, who conceptualized the phenomenon of human-place bonding (Isa et al., 2020). Tsai et al. (2020) noted that there is little evidence on place attachment and nostalgia in the hotel industry.
Over the past few decades, tourism has witnessed continued expansion and rapid growth, becoming one of the world's largest and fastest-growing economic sectors. Specifically, in 2017, tourism was the world's third-largest export category, after chemicals and fuels, ahead of automotive and food products with 1,586.00 billion USD (UNWTO, 2019). The Egyptian tourism sector has also grown to cater to a rapidly growing global tourism demand for Egyptian destinations and attractions. For example, in 2018, total revenue from tourism reached 174.10 billion EGP, a 124% increase relative to 2017, representing 15% of Egypt's GDP. The tourism sector is also one of Egypt's largest employers, offering 3.10 million jobs, or 9.5% of the total workforce (OECD, 2020). Recently, Egypt's central bank announced that the tourism sector recorded the highest level of revenues in history, reaching 13.03 billion USD in 2019, relative to 12.50 billion USD in 2010 and 11.60 billion USD in 2018 (SIS, 2020).
COVID-19 has hit the tourism sector particularly hard. For instance, the sector lost 4.2 million jobs in the Middle East, accounting for 62% of travel and tourism jobs, mainly because of declining international tourist arrivals, which dropped by 66% and are expected to reach 73% if the pandemic continues. Unsurprisingly, the Middle East's travel and tourism GDP loss in 2020 is reported to be 154.00 billion USD (WTTC, 2020). Now the world tourism industry is experiencing a critical reset. As the world fights the realities of the pandemic, there is an opportunity to reassess the future look of tourism (Brouder, 2020). For instance, in Spain, Arbulú et al. (2021) concluded that encouraging domestic tourism and reorienting outbound tourism can be a helpful strategy to relieve the travel industry crisis in the situation of significant loss of international demand. Concerning our context, to mitigate the pandemic's negative effect on the hospitality industry, Egypt aimed at reviving the tourism sector through promoting domestic tourism.
This study mainly examines the impact of servicescape and place attachment on customer engagement behaviors. Specifically, we concentrate on servicescape's direct and indirect impacts through place attachment on cognitive and affective customer engagement and on two subconstructs of behavioral engagement: willingness to suggest (WTS) and word-of-mouth (WOM). Our findings can help hospitality service providers, managers, and policymakers understand the role of servicescape in nourishing place attachment and engagement behaviors. The findings can also guide policymaking strategies regarding inbound tourism to Egypt, with the need to rethink the recovery and future of the hospitality and tourism industry.
This research contributes to the literature on many fronts. Its foremost contribution is that this study helps enrich research streams on servicescape in consumer behavior by thoroughly modeling servicescape and placing attachment with customer engagement. Hightower et al. (2002) concluded that the relationship between servicescape and marketing constructs appears to be more than bivariate. Existing research has broadly examined the mediating role of some cognitive and affective factors in the relation between servicescape and engagement behaviors, such as attitude and satisfaction (Choi & Kandampully, 2019;Line & Hanks, 2019;Peng et al., 2020;Tran et al., 2020), and value and image perceptions (Dedeoglu et al., 2018;Durna et al., 2015;Li, 2021). However, place attachment as a mediator linking servicescape with diverse forms of customer engagement behaviors has limited empirical validation.
Second, this study considerably expands the literature of servicescape-driven and attachmentdriven engagement by exploring the effects of servicescape and place attachment on four diverse and distinct constructs of customer engagement. Li (2021) argued that the investigation of customer engagement and servicescape is in demand given the limited research linking both in the same model. Indeed, the vast majority of existing research employed customer engagement as a uni-dimensional behavioral component (Dedeoglu et al., 2018;Durna et al., 2015;Hightower et al., 2002;Kim & Moon, 2009;Line & Hanks, 2019;Peng et al., 2020). However, our research conceptualizes customer engagement as a multi-dimensional variable incorporating the cognitive, affective, and behavioral constructs proposed by Brodie et al. (2011) and Hollebeek et al. (2014). Moreover, our research draws upon previous arguments that among all non-transactional outcomes of customer engagement, offering suggestions and recommendations for improvements and spreading positive WOM were considered the most prevalent behaviors (Brodie et al., 2011;Kumar & Pansari, 2016;Van Doorn et al., 2010).
Lastly, this is the first attempt to investigate the phenomenon of servicescape in consumer behavior in the Egyptian context. A great deal of research examines the role of servicescape in consumer behavior in developed economies, as in the USA (Choi & Kandampully, 2019;Hightower et al., 2002;Line & Hanks, 2019;Line et al., 2018), Canada (Kim & Moon, 2009), Spain (Tubillejas-Andrés et al., 2020), and Hong Kong (Dong & Siu, 2013). However, less research has examined the servicescape on customer perceptions and behaviors in emerging economies, where tourism and hospitality could promote economic growth and development, as in China (Peng et al., 2020;Xu & Gursoy, 2020), Turkey (Dedeoglu et al., 2018;Durna et al., 2015), Vietnam (Tran et al., 2020), Indonesia (Isa et al., 2020), and Iran (Taheri et al., 2020). There is no empirical validation on the applicability of servicescape in the Egyptian context. The remainder of the paper is structured as follows. We review the literature in Section 2. Section 3 links servicescape to customer engagement and examines the mediating effect of place attachment. Research methodology is discussed in Section 4, and our research findings are presented in Section 5. We conclude by presenting a discussion, implications, and future research directions in Section 6.

S-O-R model
The well-known Stimulus-Organism-Response (S-O-R) model, created by Mehrabian and Russell (1974), examines the effect of the environment on consumer behavior. According to Lin and Mattila (2010), the stimulus (S) is defined as an environmental factor influencing the customer's internal reactions, which represent organism (O), including the customer's emotions, opinions, beliefs, and attitudes leading to the approach or avoidance response (R). The S-O-R framework is widely adopted in hospitality research to examine the effect of the servicescape on different customer reactions and responses (Bitner, 1992;Choi & Kandampully, 2019;Dedeoglu et al., 2018;Lin & Mattila, 2010). The current study models the hotel's servicescape as the stimulus, place attachment as the organism, and customer engagement as the response.

Servicescape
Servicescape can be traced back to the term "atmospherics" introduced by Kotler (1973) to design the consumption environment in such a way that enhances the customer's emotional reactions. Bitner (1992) coined the physical surroundings in the service setting as "servicescape" and examined its impact on the consumption behaviors of employees and customers. According to Bitner (1992), servicescape is composed of three elements, ambient conditions (e.g., temperature, music, air quality, noise, etc.), spatial layout/functionality (e.g., layout, equipment, furnishings, etc.), and signs/symbols and artifacts (e.g., signage, style of décor, personal artifacts, etc.). Bitner's framework is a building block of the physical environment and is considered an essential antecedent to consumption behaviors. Tombs and McColl-Kennedy (2003) addressed that the social component may significantly influence consumers' perceptions of servicescape. The social concept has gained equal interest alongside the physical concept in the servicescape literature. According to Dong and Siu (2013), social servicescape is how the consumption environment is introduced, interpreted, and analyzed during various social interactions between customers and employees. These interactions are related to the customer evaluations of employee behavior and image and cultural elements. Reviewing the servicescape literature, one stream of studies looked at servicescape as uni-dimensional, either physical Rosenbaum & Montoya, 2007) or social Line & Hanks, 2019). Another stream examined servicescape as a bi-dimensional variable that includes both physical and social components (Dong & Siu, 2013;Al Halbusi et al., 2020;Meng & Choi, 2018). Motivated by the hospitality literature (Li, 2021;Li & Wei, 2021;Sun et al., 2021), this current study follows the second stream, considering the broader servicescape concept.

Place attachment
Place attachment represents the bond between the customer and a specific place , and that bond is due to the interactions occurring with and in that place (Lee et al., 2012). Yuksel et al. (2010) stated that place attachment is a customer's cognitive, functional, and emotional connection with a special place. The term "attachment" involves effect and emotions, while "place" includes the environmental settings where persons are cognitively and emotionally attached. The increasing place attachment improves the sense of belonging to the place (Rosenbaum & Montoya, 2007).
Prior literature (Loureiro & Sarmento, 2019;Sun et al., 2021;Williams & Vaske, 2003) suggested that place attachment consists of two dimensions: place identity and place dependence. Place identity is an emotional attachment and refers to the person's emotional, symbolic attachment toward a specific place. Conversely, place dependence is a functional attachment and refers to how a specific place matches the customer's expectations, wants, and needs. These two dimensions have been widely adopted and examined in different sectors, especially hospitality and tourism (Loureiro, 2014;Loureiro & Sarmento, 2019;Sun et al., 2021). Based on reviewing the prior literature, we argue that there is support for operationalizing the two dimensions.

Customer engagement behaviors
In business research, several perspectives conceptualize customer engagement. According to Brodie et al. (2011), the vast majority of the literature relies on a multi-dimensional perspective to customer engagement, while over 40% adopts a uni-dimensional perspective. Within the multidimensional perspective, customer engagement comprises behavioral and attitudinal (with cognitive and affective) dimensions (Brodie et al., 2011;Hollebeek et al., 2014). The definition proposed by Hollebeek et al. (2014) inclusively and comprehensively describes the constituents of customer engagement. Cognitive engagement is defined as a customer's degree of firm-related thoughts and cognitive processes during an interaction with a firm. Affective engagement refers to the level of a customer's firm-related emotions and feelings. Behavioral engagement is the degree of effort, energy, and time a customer spent during this interaction.
Although the behavioral dimension has received the most attention within the uni-dimensional perspective (Choi & Kandampully, 2019;Van Doorn et al., 2010), behavioral dimensionality is not an agreed-upon concept. For example, Van Doorn et al. (2010) described customer engagement as a behavioral aspect toward a brand or firm beyond transactional outcomes. Kumar and Pansari (2016) viewed it as the customer contributions that directly add value to a firm through revisits and repurchases, or indirectly through positive referrals, sharing knowledge, and suggestions.
According to So et al. (2020), customer engagement is a hot topic in hospitality research. Current and seemingly future studies try to uncover its predictors from diverse perspectives. They defined customer engagement as an interaction between the customer and the brand through affective, cognitive, and behavioral responses beyond purchase transactions. Therefore, the current study considers that customer engagement contains cognitive, affective, and behavioral components. However, it conceptualizes behavioral dimension using the beyond transactional outcomes from Choi and Kandampully (2019): willingness to suggest (WTS) and word-of-mouth (WOM). These two non-transactional constructs are discussed below in detail.

Willingness to suggest
As stated, customers can be behaviorally engaged through multiple non-transactional behaviors. The first form of behavioral engagement pertains to customer behaviors directed toward other customers, such as customer influences on social media through blogging, posting, rating products, and customer referrals to others. The second form pertains to customer behaviors directed toward the firm, such as customer knowledge sharing through feedback and improvement suggestions (Kumar & Pansari, 2016;Van Doorn et al., 2010). This later form was coined by Choi and Kandampully (2019) as "WTS." Though named differently, other authors described this behavior in the hospitality and tourism context. For example, Li and Wei (2021) characterized it as providing both solicited and unsolicited suggestions to hotels, which could help in enhancing service improvements. In addition, Zhang and Xu (2019) described it as the behaviors that directly support the destination, such as providing tourism-related suggestions and reporting tourists' feedback.
Encouraging complaints and suggestions can create value for the firm. Indeed, customers are uniquely positioned to offer constructive ideas to the firm, and they have service-experience perspectives that allow them to provide remarkable suggestions (Bettencourt, 1997). Moreover, Grissemann and Stokburger-Sauer (2012) argued that engaged customers are a valuable resource since they are more willing to repeat purchases, recommend the company to others, and offer suggestions that could help design a product or a service to meet customer prospects. Finally, the active and responsible customer participation behavior (e.g., exchanging information, data, and ideas with the firm and others, informing the firm about any problems that might affect performance, and suggesting remedies) is essential for improving the organizational activities (Van Dyne et al., 1994).

Word-of-mouth
As a form of non-transactional behavior, WOM is a prominent source for disseminating information about products and services (Choi & Kandampully, 2019;Yen & Tang, 2019). Prior studies by Hanks and Line (2018) and Yen and Tang (2019) stated that WOM acts as a two-edged sword that stems from customers' attitudes toward a brand. This informal behavior can be directed toward actual and potential customers (Zhang & Xu, 2019) and can happen by recommending or warning other customers from dealing with the company (Harrison-Walker, 2001). Indeed, in the hospitality context, WOM may play an essential role in shaping consumer behavior considering the difficulty of pre-consumption evaluation (Huete-Alcocer, 2017). As a voluntary behavior, engaged customers actively recommend hotels to their friends and relatives by spreading positive details about the properties (Li & Wei, 2021).

Servicescape and customer engagement behaviors
The servicescape influences customer behavior, satisfaction, and loyalty as part of the service encounter. For example, Tran et al. (2020) explored the remarkable influence of servicescape on the satisfaction and loyalty of Vietnamese coffee stores customers. In Spain, Tubillejas-Andrés et al. (2020) also found that a pleasant servicescape positively improved the satisfaction and loyalty of opera-goers. Likewise, Taheri et al. (2020) indicated that servicescape plays a significant role in curbing dissatisfaction and misbehavior among travelers in Iran. These are consistent with Hightower et al. (2002), who stated that servicescape is an essential determinant of consumer perceptions and future behaviors.
Investigating the role of servicescape on customer engagement is essential for improving performance and persuading customers (Li, 2021). The extant hospitality literature hypothesized the link between servicescape and customer engagement, with findings indicating the impact of servicescape on a variety of engagement behaviors. Kim and Moon (2009) postulated that restaurant customers could evoke cognitive, affective, and behavioral engagement in response to environmental cues. Islam et al. (2019) revealed a positive effect of physical-based and staffbased hotel quality on customer engagement. Li (2021) found that hotel customers who experience pleasant physical and social servicescapes were engaged through activating cognition, evoking summative and enduring effects, and devoting energy to the hotel. Li and Wei (2021) also found that well-designed physical and social hotel servicescape elements positively influence customers' desires to build close cognitive, affective, and behavioral connections with the hotel.
The link between servicescape and behavioral intentions is evident in the literature, indicating that servicescape can affect a variety of behavioral constructs. For example, Durna et al. (2015) indicated that hotel customers' perceptions regarding servicescape components influence their behavioral intentions (i.e., revisiting intention and WOM). The results of Hanks and Line (2018) in restaurants and Line and Hanks (2019) in hotels found that, in addition to physical servicescape, the social servicescape is also a key determinant of post-consumption behavioral intentions (i.e., revisit intention, WOM, electronic WOM). Peng et al. (2020) also demonstrated that servicescape positively impacts satisfaction and behavioral intentions to revisit and spread WOM about the integrated resort. Tran et al. (2020) concluded that servicescape elements strongly motivate or restrict the customer-to-customer and customer-to-staff knowledge sharing. Choi and Kandampully (2019) confirmed that physical and social atmospheres were crucial to guests' satisfaction and positive behavioral outcomes in upscale hotels.
Consequently, servicescape in upscale hotels and resorts is expected to facilitate customer engagement. Thus, we propose the following hypothesis: H1: There is a direct positive relationship between servicescape and cognitive engagement (H1a), affective engagement (H1b), WTS (H1c), and WOM (H1d).

Servicescape and place attachment
Environmental psychology examines the relationship between the environment and humans, and the physical environment plays a vital role in forming an individual's attachment to a place (Xu & Gursoy, 2020). Indeed, understanding attributes that can influence customer attachment in any hospitality experience seems critical. Much research has investigated the impact of the servicescape of a place on customer attachment. For example, Line et al. (2018) found that a customer perceiving the servicescape pleasantly and sympathetically results in a high level of place attachment. Isa et al. (2020) also found that the atmosphere and infrastructure of a tourism destination were positively related to social bonding and the sense of belonging and possessiveness toward the place. Xu and Gursoy (2020) revealed that the servicescape is an influential factor for generating place attachment in short-term accommodation services. Recently, Sun et al. (2021) concluded that a pleasant physical environment promotes the interactions between hotel staff and guests, influencing guests' emotions and inducing place attachment.
Given that servicescape is fundamental in forming an attachment to a place, we put forward the following hypothesis: H2: There is a direct positive relationship between servicescape and place attachment. Rosenbaum and Montoya (2007, p. 47) mentioned that place attachment is "the bonding between a person and a place". Place attachment implies the individual identifies with the physical attributes and the bonds manifested in relationships with other people (Loureiro, 2014). Rather et al. (2019) revealed a positive effect of place attachment on customer engagement: customer intentions to revisit and recommend to others. Loureiro and Sarmento (2019) mentioned that when visitors are emotionally connected to a destination, they are more engaged, share knowledge for improvements and support, spread positive referrals, and make incentive discussions.

Place attachment and customer engagement behaviors
Numerous studies have demonstrated that place attachment in hospitality and leisure settings could lead to positive behavioral engagement. For example, Loureiro (2014) concluded that the preference and satisfaction achieved from being in a place have a significant and positive effect on behavioral intentions of returning to and recommending the place. Line et al. (2018) indicated that when restaurant patrons felt strongly attached, they were more inclined to tell others about the restaurant. Isa et al. (2020) posited that the place attachment dimensions have a significant and positive influence on customer behavioral intentions. Tsai et al. (2020) also concluded that place attachment among customers in the hotel industry result in favorable behavioral intentions of recommendation to others and revisits.
Based on this body of literature across hospitality and tourism contexts, we propose the following hypothesis: H3: There is a direct positive relationship between place attachment and cognitive engagement (H3a), affective engagement (H3b), WTS (H3c), and WOM (H3d). Bitner (1992) argued that cognitive and affective factors mediate servicescape's impact, and Hightower et al. (2002) recommended examining a model beyond the bivariate relationship between servicescape and customers' future behaviors. Several studies have further advanced this research agenda.

Mediating role of place attachment
At the outset, Kim and Moon (2009) posited that the stimuli from the servicescape indirectly influence customer behaviors through pleasure-feeling and perceived service quality. Durna et al. (2015) demonstrated the indirect effect of servicescape components on customer behavioral intentions through customer perceptions in the hotel context. Dedeoglu et al. (2018) stated that emotional responses to physical and social stimuli could mediate between the environment and post-consumption behaviors. Line et al. (2018) posited that delighted and kind servicescape elements affect consumer emotional responses, making the customer feel happy, welcome, and at home. Tsai et al. (2020) identified place attachment as a mediator in the relationship between experiential quality and behavioral intentions. Loureiro and Sarmento (2019) argued that place-attached visitors tend to become more active and interactive in providing improvement suggestions, promoting, and discussing the benefits of visiting the place. Li and Wei (2021) reinforced the premise that physical and social servicescapes can indirectly drive customers to offer suggestions and recommend the hotel to others. Isa et al. (2020) posited that the environment is a stimulus to place attachment that ultimately determines customer behavioral intentions.
Over and above, prior studies applied the S-O-R paradigm to test the relationships between service evaluation, consumption-related emotions, and behavioral outcomes (Bitner, 1992;Choi & Kandampully, 2019;Dedeoglu et al., 2018;Sun et al., 2021). When modeled in the S-O-R framework, the servicescape acts as an environmental factor (stimuli), place attachment reflects individuals' psychological state (organism), and the components of engagement serve as behavioral responses (responses).
To investigate how place attachment mediates the relationship between servicescape and customer engagement behaviors, we put forward the following hypothesis: H4: Place attachment mediates the relationship between servicescape and cognitive engagement (H4a), affective engagement (H4b), WTS (H4c), and WOM (H4d).
Based on this review and the hypothesized direct and indirect servicescape-customer engagement relationships, Figure 1 illustrates our research model.

Data collection and sampling
Egypt is the most populous country in the Middle East and the third in Africa with 104 million people and is growing at a rate of 2.28% with an urban population of around 43 percent of the total population and an urbanization rate of 1.86% annually. Egypt is one of the few countries that had a positive growth rate in 2020, and Egyptian GDP is estimated to hit the ceiling of 400 USD billion in the year 2021 (Santander Trade, 2021). We used Cochran's formula for infinite population for estimating sample size. The formula is SS = [Z 2 p (1 − p)]/ C2 (Godden, 2004). The confidence intervals (C) are 0.05, the confidence level is at 95 percent (z value is 1.96) and percentage of population is (p) is 0.5. This means 385 surveys are adequate to have a confidence level of 95% that the real values are within ±5% of the surveyed values.
Following Choi and Kandampully (2019) and Line and Hanks (2019), the present study has prescribed the upscale hotels and resorts to be investigated, thus, we argue that such a limit help exploring the research constructs. It is impossible to directly select a sample from the population in Egypt, as there is no list of customers' names who visited hotels in the last three years. So, convenience sampling (non-random sampling) has been determined as the most suitable technique to collect data where the target population members have met specific criteria, such as, geographical proximity, easy accessibility to the researcher, and willingness to participate (Etikan, 2016).
The constructs were measured using Google Forms, and the questionnaire's link was posted on Facebook. Google Forms was chosen for several reasons; first, it is useful when the sample is largely demographically diverse. Second, hotel customers in Egypt are widespread. Third, as mentioned, there is no list of customer names who visited hotels in the last three years, so it is difficult to select a sample from the population. To ensure a consistent frame of reference within the sample, respondents were asked to participate if they stayed in an upscale hotel or resort over the past three years.
The sampling method began as a convenience sample, which developed into snowball sampling as more respondents recommended other participants. The time needed to complete the survey is approximately 10-15 min. In total, 411 questionnaires were completed. A number of 6 responses were identified as multivariate outliers and then removed. This resulted in a final total of 405 valid responses which is considered sufficient for the analysis. Table 1 shows our sample profile.
As shown in Table 1, around 66% of the respondents were female, and 34% were male. This ratio is close to these prior studies (e.g., Li & Wei, 2021;Taheri et al., 2020;Xu & Gursoy, 2020). Egyptian visitors of upscale hotels and resorts in our sample were 35-44 years old (37%), 25-34, and 18-24 with ratios of 29% and 25%, respectively. These ratios are also close to these prior Xu & Gursoy, 2020) also reported nearby ratios. Most of the participants (85%) visited an upscale hotel or resort over the past three years three times or less, and 40% of them were new customers with just one visit. According to the length of stay, the three nights' stay ranked first with 33%, followed by two nights with 25%. The majority of respondents stayed for three nights or less. Choi and Kandampully (2019) reported that more than 80% of visits were three nights or less. In addition, Li and Wei (2021) revealed that the length of stay of 75% of participants was one or two nights. Nearly three fourth of participants attributed the reason for the visit to leisure and recreation, which indicates the pivotal role of nature and leisure settings in maintaining supportive customer behaviors, which was the same as disclosed by Choi and Kandampully (2019).

Questionnaire and measures
For collecting data, we used an online survey. The questionnaire included items on the dimensions of servicescape, place attachment, and customer engagement (cognitive, affective, and WTS and WOM as the behavioral dimension). All variables were measured using multiple items with a 5-point Likert scale. We measured demographic variables by asking about gender, age, education, and occupation. We also asked questions about; repeat visits over the last three years, the number of nights stayed per visit, and the visit's reason.

Results
In structural equation modeling, two main steps were conducted to assess the proposed model. First, the measurement model was utilized to evaluate the reliability and validity of the study's variables. Second, the structural model was conducted to examine the model fit and test the hypothesis (Anderson & Gerbing, 1988).

Preliminary and descriptive analysis
We conducted the confirmatory factor analysis (CFA). The results of CFA depicted in Table 2 show that items SC1, SC6, SC10, and SC21 were removed from the servicescape construct, and the item PA10 was removed from the construct of the place attachment because their factor loadings are less than 0.6. Factor loadings were used to assess the indicator reliability. Looking at Table 2, the estimated factor loadings for all items were 0.6 or higher. Hence, the model has met the necessary conditions for reliability (Hair et al., 2010). Cronbach's alpha was utilized to assess the internal consistency. Table 2 shows that all Cronbach's alpha scores exceeded the 0.70 guideline (Hair et al., 2010). Furthermore, Table 2 shows the mean and standard deviation of our constructs. The descriptive statistics (mean and standard deviation) of the second-order constructs were servicescape (4.356 and 0.511) and place attachment (3.819 and 0.687). The mean and standard deviation of the first-order constructs were cognitive engagement (3.45 and 0.871), affective engagement (4.090 and 0.651), willingness to suggest (4.140 and 0.623), and (4.320 and 0.617) of the word-ofmouth.

Measurement model validation
The psychometric properties of the questionnaire were assessed by testing its reliability and validity. According to Nunnally and Bernstein (1994), the construct reliability is examined using (a) composite reliability (CR) and (b) Cronbach's alpha. As shown in Table 2, Cronbach's alpha scores exceed the required threshold of 0.70 for all our model constructs. Construct reliability is also confirmed with all CR values exceeding 0.70, as Table 3 portrays.
According to Henseler et al. (2009), two validity sub-types should be examined: (a) the convergent validity and (b) the discriminant validity. The average variance extracted (AVE) was tested as a criterion for the convergent validity, as recommended by Hair et al. (2010). The value of all AVEs presented in Table 3 exceeded the minimum criterion of 0.50, suggesting that our model constructs are convergently valid. It's worth noting that for all constructs, CR was found to be greater than AVE, and both are found to be higher than 0.50. These values, combined with the  significant factor loadings shown in Table 2, provide strong evidence of convergent validity of our model constructs as presented by Line et al. (2018). Discriminant validity was assessed using the Fornell-Larcke criterion (Fornell & Larcker, 1981). As presented in Table 4, the square root of the AVE, as shown along the diagonal of each latent variable, was greater than its correlation with the other latent variables, confirming the discriminant validity of our model constructs. Therefore, drawing on the measurement model analysis, all the model constructs are reliable and statistically robust.

Structural model and hypotheses testing
To test the research hypotheses, the structural equation model approach (SEM), and bootstrapping was used for the mediation analysis with AMOS 23.0. Results of path coefficients are presented in Table 5. The model fitness investigates how the model estimates the correlation between variables in the data (Hair et al., 2010). The goodness-of-fit indicators of the structural model are examined by drawing on the X 2 À statistics, the traditional measure for evaluating overall model fit, Comparative Fit Index (CFI), Incremental Fit Index (IFI), Tucker-Lewis Index (TLI), and the Root Mean Square Error of Approximation (RMSEA; Hu & Bentler, 1999).
SEM yielded the following results: X^2/df = 1530.52/713 = 2.147, p < .001, which is an acceptable ratio for this statistic (i.e., it's within the range of the low as 2.0 to the high as 5.0). The value of CFI = 0.917, which is greater than 0.90, to ensure the model is not misspecified. IFI = 0.918, with a value close to 1.0. TLI = 0.904, which is greater than the threshold of 0.90 for the model to be considered a satisfactory fit. RMSEA = 0.061 is greater than the cut-off value close to 0.06 and less than the stringent upper limit of 0.07. The model fit is deemed acceptable, indicating the appropriateness of the structural model as indicated by previous studies (e.g., Forza & Filippini, 1998;Hair et al., 2010;Hu & Bentler, 1999).
The results depicted in Table 5 confirmed that servicescape (SC) was a significant positive predictor of most of the customer engagement behaviors, partially supporting H1. Particularly, among the four engagement constructs, cognitive engagement (CE) was not significantly related to SC, with path coefficient β = -0.100 (p = 0.052), so H1a is not supported. This result is inconsistent with Bitner's (1992) argument that the servicescape may directly elicit cognitive responses. This result also contradicts the recent findings by Li (2021), who argued that when Chinese hotel customers receive a pleasant environment, they tend to activate hotel-related thought processing (cognitively engagement), which implies the unique role of the hotel setting. Whereas SC had a significant and positive relationship with affective engagement (AE), with path coefficient β = 0.126 (p < .01), supporting H1b. This result is similar as reported by prior studies (e.g., Kim & Moon, 2009;Li, 2021). SC had a significant and positive relationship with WTS, with path coefficient β = 0.152 (p < .05), supporting H1c. The relationship between SC and WOM was positive and significant, with a path coefficient β = 0.144 (p < .01), supporting H1d. Our results regarding the role of servicescape in behavioral intentions are in line with previous studies (e.g., Kim & Moon, 2009;Li, 2021;Line & Hanks, 2019).
The relationship between SC and place attachment (PA) was significant and positive, with path coefficient β = 0.568 (p < .01), supporting H2. This result is consistent with previous studies in the hospitality sector (e.g., Line et al., 2018;Sun et al., 2021;Xu & Gursoy, 2020). In support of H3, PA significantly influenced all the customer engagement behaviors since all paths were significant. In details, PA had a significant and positive relationship with CE, with path coefficient β = 0.716 (p < .01), supporting H3a. The relationship between PA and AE was positive and significant, with path coefficient β = 0.653 (p < .01), supporting H3b. These findings of the relationship between place attachment and customer engagement are in line with prior literature (e.g., Rather et al., 2019). The positive relationship between PA and WTS was also significant, with path coefficient β = 0.384 (p < .01), supporting H3c. PA was also significantly related to WOM, with path coefficient β = 0.534 (p < .01), thus, H3d is also supported. These results regarding the role of place attachment in behavioral engagement are in line with previous studies (e.g., Loureiro, 2014;Loureiro & Sarmento, 2019;Rather et al., 2019;Tsai et al., 2020).
The mediating effects were assessed by applying the bootstrapping procedure suggested by Jose (2013), with a confidence level of 95% and a bootstrap sample of 2,000. We used the biascorrected bootstrapping approach approached by Demming et al. (2017) for many reasons; it is suitable for larger sample sizes, helps in adjusting the confidence interval bias in the bootstrap sample distribution, and is the best choice for testing the mediation analysis. According to Table 4, results asserted the significant indirect effects of servicescape on all customer engagement behaviors through place attachment, supporting H4. The mediating impacts were evaluated using the variance accounted for (VAF) approach developed by Hair et al. (2013). The VAF is the proportion of the total effect of a construct because of indirect effects. If the VAF is more than (80) thus, there is a full mediation effect. If the VAF score is less than (20) thus, there is no mediation effect. If the VAF score falls between (20 and 80) thus, there is a partial mediation effect.
For illustration, PA plays a fully mediating effect in the relationship between SC and CE, thus supporting H4a. The coefficient for the total effect of SC on CE was 0.307, and the coefficient for the direct effect of SC on CE was −0.100. The indirect effect of SC on CE was 0.407; 95% CI (0.309, 0.502). These results demonstrated that the total effect of SC on CE through mediating PA was 132.57 percent (0.407/ 0.307). So, the place attachment has fully mediated the relationship between the servicescape and cognitive engagement behavior.
Additionally, PA plays a partial mediation effect in the relationship between SC and AE, WTS, and WOM, respectively, which supports H4b, H4c, and H4d. The coefficient for the total effect of SC on AE was 0.497, and the coefficient for the direct effects of SC on AE was 0.126. The indirect effect of SC on AE was 0.371; 95% CI (0.276, 0.459). These results demonstrated that the total effect of SC on AE, when mediated by PA, was 74.64 percent (0.371/0.497). Similarly, PA has also partially mediated the relationship between SC and the two types of customer's behavioral engagement of WTS and WOM. For illustration, the coefficient for the total effect of SC on WTS was 0.370, the coefficient for the direct effect of SC on WTS was 0.152. The indirect of SC on WTS was 0.218; 95% CI (0.147, 0.292). The total effect of SC on WTS mediated by PA was 58.91 percent (0.218 /0.370). The coefficient for the total effect of SC on WOM was 0.448, the coefficient for the direct effect of SC on WOM was 0.144. The indirect of SC on WOM was 0.304; 95% CI (0.228, 0.388). These results demonstrated that the total effect of SC on WOM, when mediated by PA, was 67.85 percent (0.304 /0.448). Thus, these engagement behaviors (AE, WTS, and WOM) are explained by direct and indirect effects of servicescape through place attachment.
This study's findings assert that place attachment can be appealed to promote the influence of servicescape on customer engagement. This is in line with Xu and Gursoy (2020), who revealed the mediating role of place attachment in the relationship between servicescape elements and intention to recommend. These results also support the early researchers' conclusions regarding the mediating role of cognitive and affective constructs in the relation between servicescape attributes and consequences of the customer experience. For example, Hightower et al. (2002) concluded that servicescape significantly influences behavioral intentions; however, various marketing constructs mediate this effect. Additionally, Isa et al. (2020) suggested the path effect of environmental cues on place attachment and revisit intention.

Discussion
The authors developed the proposed model to investigate the impact of servicescape on customer engagement behaviors mediated through the place attachment in upscale hotels and resorts in Egypt.
This study shows that servicescape positively relates to customer engagement behaviors. Specifically, the findings support the direct relation between servicescape and affective engagement and the two sub-constructs of behavioral engagement: WTS and WOM. In contrast, the results do not support a direct relationship between servicescape and cognitive engagement. The findings suggest that when customers perceive a pleasant servicescape in the hotel or resort, they affectively engage, make suggestions for improvements, and make positive recommendations. Overall, these findings support Li and Wei (2021): a hotel's well-designed servicescape is essential in influencing customer engagement. In their works, Hanks and Line (2018) and Line and Hanks (2019) stated that environmental, physical, and social cues significantly affect the customer's experience assessment and post-consumption behaviors.
We also found that place attachment positively relates to customer engagement behaviors. Place attachment, composed of place identity and place dependence, may drive customer cognitive and affective engagement behaviors, such as making suggestions for improvements and positive recommendations. In their study, Loureiro and Sarmento (2019) argued that tourists who feel inner and social identification with a place would be emotionally attracted and more motivated to engage.
Our model indicates that place attachment mediates the hypothesized relationship between servicescape and customer engagement behaviors. The findings show that customers become more attached to the hotel with a pleasant servicescape, with a higher inclination toward engagement behaviors, including cognitive engagement. In particular, they seem to develop identification and dependence toward the hotel, thereby increasing their engagement forms. One possible explanation is based on the attachment theory, wherein the favorable emotional attachment process may generate positive cognitive evaluation, representing an impetus for action and interaction (Loureiro & Sarmento, 2019).
Interestingly, Choi and Kandampully (2019) showed that while satisfied customers were more likely to share positive remarks about the hotel with others, they were less likely to provide suggestions for improvements. This might pose a question about the post-consumption behaviors exerted by satisfied and attached customers. Our findings imply that customers respond to physical cues and social interactions so that their responses are proportionate to their place attachment levels. In this regard, an attached customer looking at maintaining a long-lasting relationship with the destination may lead to the positive association between place attachment and WTS; thus, such customers have more to suggest. Furthermore, from the descriptive statistics, servicescape has the highest mean score of 4.356. Therefore, it can be assumed that respondents received an enjoyable servicescape and felt respected, valued, and well-treated.
In summation, our results revealed that servicescape directly influences place attachment and customer engagement, and place attachment plays a critical role in bracing customer engagement. Generally speaking, place attachment mediated the relation between servicescape and customer engagement dimensions, including two prominent forms of behavioral engagement.

Theoretical implications
This study provides several theoretical implications. First, the research responds to the call from Al Halbusi et al. (2020) and Li and Wei (2021) to investigate the servicescape as a bi-dimensional variable consisting of both physical and social surroundings. This thorough investigation helps discuss the consumption environment and predicts consumer behavior. Moreover, this research represents a significant empirical move in servicescape literature by addressing the significance of capturing physical and social servicescapes in hotels to affect place attachment and customer engagement behaviors. Second, this study broadens the current knowledge on the role of servicescape dimensions in enhancing customers' feelings of belonging. Line et al. (2018) and Su et al. (2018) stated that much research has focused on physical servicescape and ignored the role of social surroundings on place attachment. In contrast, the present study argues the influence of social and physical servicescapes on consumer feelings and emotions.
Third, this study substantially broadens the current knowledge on customer engagement literature by exploring its different dimensions: cognitive, affective, and behavioral. The behavioral dimension was studied by concentrating on two non-transactional outcomes. Compared with previous studies (Durna et al., 2015;Line & Hanks, 2019;Line et al., 2018) that focused on customer-to-customer interactions, the current study is a pioneering attempt to examine behavioral engagement through two types of interaction: customer-to-customer (WOM) and customer to firm (WTS). This study's findings enrich and support the framework that Choi and Kandampully (2019) asserted, which included the customers' WTS improvements to the hotel. Finally, this study is among the few studies to assess the validity of place attachment as a mediator in the relationship between servicescape elements and engagement behaviors. Xu and Gursoy (2020) examined the role of place attachment in the relationship between servicescape elements and intention to recommend. Conversely, this study enriches the servicescape literature by examining the role of place attachment regarding the effects of servicescape on a group of diverse behaviors of customer engagement.

Practical implications
Our findings provide some practical implications for the hotel industry. First, they support the positive effect of the hotel servicescape encompassing physical and social cues on place attachment and diverse forms of customer engagement behaviors. Thus, hotels should provide customers with an exceptional physical servicescape by offering them a cheerful atmosphere, attractive architecture, elegant decoration, up-to-date and well-maintained amenities, welcoming and relaxing background music, pleasant smell, attractive colors, and well-furnished rooms. To improve social servicescape, managers should train the front-of-house employees to deal with customers politely and pleasantly and induce hotel staff to adopt a customer-oriented approach with customer convenience as a top priority. A hotel servicescape should provide a memorable and meaningful experience that reinforces customers' feelings of belongingness and unleashes customer engagement behaviors.
Additionally, these findings show that highly attached customers tend to be more behaviorally engaged by providing improvement suggestions and spreading positive recommendations about the hotel. Therefore, marketing managers should design and administer surveys to determine the highly attached customers, encouraging posts on the hotel's website regarding the physical and social servicescapes. Such customers would also be inclined to share their experiences on social media such that potential guests are encouraged to visit the hotel. Furthermore, managers should encourage customers' feedback by welcoming their suggestions for improvements and designing a compensation system, such that the best suggestion receives discounts or a free stay. Hotels could also encourage highly attached customers to share the hotel's website, pages, and accounts on their social media (e.g., Facebook, Instagram, Twitter) by designing contests for specific periods and asking customers to make a particular number of likes and shares for obtaining free nights at the hotel or discounts for future stays. Lastly, they could provide compensation and offers for customers who spread positive WOM about the hotel, encouraging their friends, relatives, and colleagues to visit.

Limitations and future directions
This study has some limitations offer guidelines for future research. First, this research was conducted in solely one country, thus the generalization of results may be questionable. Future attempts might consider a multi-cultural and/or multi-country study to understand the moderating role of culture.
Second, our proposed model was examined in the upscale hotels, this may hinder the generalization of results to other classes of hotels. Third, this study was executed in the hotels, so future studies could employ our model in different settings (e.g., private and community universities, filling stations, gyms, cafes, and tourism destinations). Fourth, the hypotheses were tested using cross-sectional data; future studies can develop a longitudinal study. Fifth, the convenience sampling method was used; future studies can apply random sampling. Sixth, future studies can explore the relationship between the hotel servicescape and other variables, such as customer's perceived experience, perceived image, and citizenship behaviors. Finally, future studies may examine the moderation role of the visit purpose, demographics, the number of visits, and the stay duration.