International Journal of Hospitality Management 53 (2016) 116–128

Contents lists available at ScienceDirect

International Journal of Hospitality Management

journal homepage: www.elsevier.com/locate/ijhosman

Effect of servicescape on customer behavioral intentions: Moderating

roles of service climate and employee engagement
Kuo-Chien Chang
Department of Leisure and Recreation Management, Chihlee University of Technology, No. 313, Sec. 1, Wunhua Road, Banciao District, New Taipei City
22050, Taiwan, ROC

a r t i c l e i n f o a b s t r a c t

Article history: Although customer experience and employee behavior are essential to a company’s success, few
Received 16 April 2015 hospitality field studies have used the service marketing triangle model to empirically examine ser-
Received in revised form 5 December 2015 vice climate and employee engagement as firm-level moderators of the individual-level stimulus
Accepted 7 December 2015
(servicescape)–organism (customer emotions)–response (behavioral intentions) relationship. Data were
collected from customers, managerial employees, and employees of hot spring resorts. To account for
Keywords:
nesting effects, hierarchical linear modeling tested hypothesized relationships among variables. The
Resort
results show employee behaviors contribute to service delivery and clarify why (cross-level modera-
Servicescape
Customer emotion
tion of service climate and employee engagement at the firm level) and how (perceived servicescape and
Behavioral intention customer emotions at the individual level) of their interactions crucially influence customer consumption
Service climate experiences in a service firm. The study extends the S–O–R by demonstrating outcome variables are not
Employee engagement limited to individual-level effects, as environmental characteristics (firm-level explanatory variables) are
Multilevel modeling also major considerations.
© 2015 Elsevier Ltd. All rights reserved.

1. Introduction Rossiter, 1982; Mehrabian and Russell, 1974) to understand the

benefits associated with servicescapes. For example, in the hos-
Services, measured by GDP, currently outperform manufac- pitality field, servicescapes directly affect customer emotions and
turing in many economies, inspiring environmental psychology produce predictable customer behavioral intentions (e.g., Jani and
and marketing researchers to clarify how physical environment Han, 2015; Lin and Mattila, 2010).
influences customer service experiences (Mari and Poggesi, 2013). Although Bitner (1992) primarily focused on servicescapes’
Servicescapes (Bitner, 1992) help researchers understand the phys- physical characteristics, Arnould et al. (1998) noted that both
ical environment’s role in consumer service evaluations, which is substantive (functional clues) and communicative (human clues)
important in leisure service settings (Wakefield and Blodgett, 1994) stagings of the servicescape influence customer experiences. Thus,
and more important in hedonically motivated services (Reimer and service environments should be examined in terms of physical
Kuehn, 2005), such as in the hospitality industry (Lin, 2004). aspects and social interactions between customers and employ-
In marketing, a servicescape is defined as the landscape where ees, as these affect the overall atmosphere (Nilsson and Ballantyne,
services are experienced (Bitner, 1992). This term has been used 2014). There are concerns about service employees’ role in the
to describe the service setting’s physical aspects contributing to servicescape, corresponding to the service profit chain concept
customers’ overall physical perception of the experience (Ford (Heskett et al., 1994), suggesting there are critical relations among
et al., 2012). Therefore, suitably designed servicescapes manifest employee attitudes, customer satisfaction and, ultimately, organi-
as tasteful tangible facilities – appealing interior designs, com- zational profits.
fortable lighting and sounds, and unique fragrances – inviting Research on service employees’ role in the S–O–R relationship
consumers into constructed atmospheres satisfying needs and primarily examines the impact of service employee performance
expectations (Reimer and Kuehn, 2005). Researchers have applied on customers at the individual level (e.g., Dong and Siu, 2013; Lin
the stimulus–organism–response (S–O–R) paradigm (Donovan and and Mattila, 2010). This explains the how but ignores the why,
which together are the underlying reasons that service employ-
ees treat customers differently at the firm level. That is, when
service profit chains exposes the importance of service employ-
E-mail address: kcchang@mail.chihlee.edu.tw ees in influencing customer behaviors, employee behaviors can be

http://dx.doi.org/10.1016/j.ijhm.2015.12.003
0278-4319/© 2015 Elsevier Ltd. All rights reserved.
 K.-C. Chang / International Journal of Hospitality Management 53 (2016) 116–128 117

examined at the level of direct interactions with customers (the and Bitner, 2003). Employee engagement, meaning employees
how, at the individual level), and at the level that observes their have high energy levels, enthusiasm, helpful and service-oriented
reasons for the interaction behaviors (the why, at the firm level). behaviors, and innovative behaviors to improve service quality
The latter includes organizations’ environmental characteristics (Slåtten and Mehmetoglu, 2011), reflects external marketing in
(e.g., service climate and employee engagement), which are collec- practice.
tive and shared phenomena experienced by employees nested in The S–O–R relationship (Mehrabian and Russell, 1974) clarifies
work groups which in turn are nested in functions, nested within the relation between servicescape, emotions, and customer behav-
organizations (Salanova et al., 2005). The present study clarifies ior (Bitner, 1992) in response to interactive marketing activities
whether S–O–R relationships between individual-level variables within the service marketing triangle model. Thus, the S–O–R rela-
(i.e., perceived servicescape, customer emotions, and behavioral tionship is an interactive marketing activity through which a firm
intentions) change as a function of higher-level moderator vari- fulfills promises to customers. When the service is produced, con-
ables (i.e., service climate and employee engagement), reflecting a sumed, and satisfies the customer’s expectations, the promise is
cross-level interaction effect (Kozlowski and Klein, 2000). fulfilled (Bitner, 1995). Bitner (1992) proposed that an individual’s
Individual-level S–O–R relationships and firm-level environ- response often depends on situational factors (e.g., service climate
mental characteristics can be synthesized with the service and employee engagement) affecting an individual’s expectations
marketing triangle model (Bitner, 1995; Zeithaml and Bitner, 2003). of and feelings about an environment. Consequently, customers
At the bottom of the model, interactive marketing corresponds to experience how, at the individual level, and why, at the firm level,
individual-level S–O–R relationships, while internal and external employees deliver services. Then, they positively report their emo-
marketing at the model’s sides correspond to firm-level environ- tional responses and behavioral intentions.
mental characteristics. Thus, this study incorporates firm-level The conceptual model’s key constructs are reviewed and pro-
variables, like service climate and employee engagement (Salanova posed relationships are discussed below.
et al., 2005), with internal and external marketing to clarify their
effects on individual-level relationships of the servicescape, emo- 2.2. Servicescape
tional reactions, and behavioral responses.
Wieseke et al. (2008) noted marketing researchers tend to Bitner (1992, p. 58) describes servicescape as “the built envi-
emphasize either micro- or macro-level perspectives without rec- ronment (i.e., manmade, physical surroundings rather than natural
ognizing interactions between the levels or the importance of or social environment) affecting consumers and employees in ser-
multilevel approaches in marketing and management research. vice organizations.” Wakefield and Blodgett (1994) argued that
Thus, multilevel research is needed to provide practitioners a in leisure services (e.g., upscale restaurants or resorts) where
more accurate account of the organizational level to facilitate consumers spend longer periods, the servicescape is more likely
performance improvement. In the hospitality field, empirical to influence attitudes toward service provision. Thus, service
research is lacking on cross-level effects of service climate and providers should recognize its important role in marketing strate-
employee engagement on individual-level relationships among gies.
the servicescape, emotional reactions, and behavioral responses. Leisure-related services are often associated with the natural
Accordingly, this study applies an integrated model for prelimi- environment. Fredman et al. (2012) proposed a two-dimensional
nary investigation using hot spring resorts as the case example. nature-based servicescape model (naturalness vs. facilities and
To address different units of analysis at individual and firm levels open access vs. exclusive rights) to help researchers further
across organizations, it employs hierarchical linear modeling (HLM) understand human-nature relationships in a servicescape. Natu-
for a cross-level analysis. The results provide practical suggestions ral environments, like forests, mountains, and rivers, are important
for hot spring managers to help them build a service advantage for service providers in the nature-facilities continuum. However,
through improved servicescapes and employee behaviors. facilities with cabins and sports facilities may be more important
for activities such as hunting and fishing (Fredman et al., 2012).
Therefore, on the nature-facilities continuum, service providers’
2. Research model and hypotheses dependence on natural or constructed environments reflects the
demand side of customers’ activity preferences. For example,
2.1. Conceptual model although hot spring resorts are mostly in natural hot spring areas,
customers demand more from facilities when they want to enjoy
This study (see Fig. 1) is based on the service marketing trian- hot spring baths and spa-related activities; the natural setting is
gle model with corresponding service promises at the firm level incidental or plays a subordinate role (Valentine, 1992).
(Bitner, 1995; Zeithaml and Bitner, 2003). The S–O–R relationship Bagozzi (1975) proposed that most marketplace exchanges are
is also included at the individual level (Bitner, 1992; Mehrabian mixed, including tangible and intangible entities, in which con-
and Russell, 1974). In this model, employee behaviors from internal sumers fulfill both utilitarian and social and psychological needs.
and external marketing perspectives are treated as firm-level vari- Thus, in addition to physical stimuli, social stimuli, like interactions
ables to explore individual-level effects on interactive marketing with employees, influence customer service experiences (Bitner,
variables. 1992; Nilsson and Ballantyne, 2014; Rosenbaum and Massiah,
Internal marketing activities are crucial for enabling promises. 2011). Accordingly, the present study explores effects of substan-
Promises are easy to make, but won’t be fulfilled unless employees tive and communicative stagings in servicescapes (Arnould et al.,
are recruited, trained, provided with tools and appropriate internal 1998; Dong and Siu, 2013) on customer emotions.
systems, and rewarded for good service (Bitner, 1995). Thus, the
service climate reflects employees’ shared perceptions and use of 2.3. Customer emotions
policies, practices, and procedures that reward, support, and expect
customer service (Schneider et al., 2002), represents the spirit of Service environments are key to service delivery because they
internal marketing activity. can foster pleasant emotional reactions (Bitner, 1992). Consump-
External marketing activities also seek to make promises of expec- tion experiences, specifically for hedonic products/services, reflect
tations and delivery to customers (Bitner, 1995), and reliability in utilitarian values as well as the pleasure derived from the con-
marketing is only fulfilled when promises are delivered (Zeithaml sumption process (Jani and Han, 2015). Emotions are “positive or
118 K.-C. Chang / International Journal of Hospitality Management 53 (2016) 116–128

Fig. 1. The research model.

negative reactions or mental stages of readiness that arise as a includes four facets: global service climate, customer orientation,
consequence of specific events or circumstances” (Bagozzi et al., managerial practices, and customer feedback (Schneider et al.,
1999, p. 1). In the M-R model, bipolar scales (e.g., pleasure, arousal, 1998, 2002). Global service climate is a summary measure of the
and dominance) are empirically superior to other emotional scales organization’s service climate. Schneider et al. (1998) examined
due to ease of application (Mehrabian and Russell, 1974). However, contributions of the other three dimensions of service climate
Lo and Wu (2014) suggested a unipolar scale instead, because the to global service climate, and found they underlie global service
absolute opposite of each emotional description can be difficult to climate and combine into actual service climate as a function of
identify (Bagozzi et al., 1999). Consistent with previous research three specific sets of service practices. Accordingly, the present
(Hennig-Thurau et al., 2006; Lin and Liang, 2011), customer emo- study used the global service climate scale (Salanova et al., 2005)
tions (e.g., elated, peppy, enthusiastic, and excited) are evaluated to explore global service climate’s effect on customer emotions
using unipolar scale exploring effects of customer emotions on and the servicescape–customer emotions relationship.
behavioral intentions.
2.6. Employee engagement
2.4. Behavioral intentions
Kahn (1990) was the first to discuss employee engagement
Behavioral intentions are “the degree to which a person has for- (Slåtten and Mehmetoglu, 2011), defining it as “simultaneous
mulated conscious plans to perform . . . future behavior” (Warshaw employment and expression of a person’s preferred self in task
and Davis, 1985, p. 214). Zeithaml et al. (1996) suggested that favor- behaviors that promote connections to work and to others, personal
able behavioral intentions indicate that customers have bonded presence (physical, cognitive, and emotional), and active full role
with the company, and include positive reports about it, rec- performances” (p. 700). Although some (e.g., May et al., 2004) view
ommending services to others, and paying a price premium and employee engagement as multifaceted, comprising two or more
retaining cognitive loyalty to it. Behavioral intentions are a surro- separate components, Britt et al. (2007) argued several dimensions
gate indicator of actual behavior (Fishbein and Ajzen, 1975). Thus, are antecedents or outcomes of engagement. Britt (1999) proposed
when customers express return intentions and recommend ser- a single scale of self-engagement in work, defined as “individuals
vices to others (Kuo et al., 2012), these behaviors are viewed as feeling a sense of responsibility for and commitment to a perfor-
outcomes influenced by customer emotions. mance domain so that performance matters to the individual” (Britt
et al., 2005, p. 1476). Accordingly, the present study uses employee
engagement, defined as self-engagement in work (Britt, 1999; Britt
2.5. Service climate
et al., 2005), to explore its effect on both behavioral intentions and
the customer emotions–behavioral intentions relationship.
Schneider et al. (1980) were among the first to investigate
“service climate,” which emerged from the generic construct of
organizational climate (Cha and Borchgrevink, 2014). A firm’s 2.7. Research hypotheses
service climate can be defined as “employee perceptions of the
practices, procedures, and behaviors that get rewarded, sup- The service delivery process comprises two levels. At the indi-
ported, and expected concerning customer service and customer vidual level, employees interact with customers, while the firm
service quality” (Schneider et al., 1998, p. 151). Service climate level comprises environmental characteristics resulting from the
 K.-C. Chang / International Journal of Hospitality Management 53 (2016) 116–128 119

firms’ philosophy, attitude, and strategies (Karantinou and Hogg, communicative and substantive staging have equal effects on
2001). As Fig. 1 shows, both cross-level direct effects and cross-level customers’ pleasure levels. Accordingly, the present study used
moderating effects are testable with multilevel modeling. Cross- individual-level substantive and communicative staging as per-
level direct effects highlight effects of global, shared, or configural ceived servicescapes to explore how these interact with firm-level
unit properties on lower-level constructs. Cross-level moderating service climate, which also affects customer emotions.
effects reveal whether the relationship between two lower-level Improving service climate is a concrete means for corporations
constructs is changed or moderated through a higher-level charac- to enhance employee’s service behavior (Schneider et al., 1998). A
teristic in which both are embedded (Kozlowski and Klein, 2000). positive service climate motivates employees to perform because
the cues received from the work environment signal that this
2.7.1. Cross-level direct effects: firm-level service climate’s effect behavior is valuable to the organization and customers (Schneider
on individual-level customer emotions and Barbera, 2014). Lin (2004) examined servicescape design ele-
Climate—the “feeling in the air”—is the “atmosphere that ments’ effects on consumer responses in the hospitality field and
employees perceive is created in their organizations by practices, discovered that intervening variables, e.g., how service climate
procedures, and rewards” (Schneider et al., 1994, p. 18). Thus, ser- affects individual-level perceptions, should be considered from a
vice climate is a collective and shared phenomenon created through macro-perspective.
organizational practices focused on customer service (Salanova Lewin’s (1951) field theory suggests individuals’ behavior
et al., 2005). A climate for service tells employees superior cus- resulted from an interplay between personal characteristics
tomer service is expected and rewarded. Employees are thus more (e.g., ability) and the psychological environment (e.g., how the
likely to provide good service to meet customer needs (Liao and environment is perceived) (Manning et al., 2005). Thus, the ratio-
Chuang, 2004; Wang, 2015). nale for relating firm-level service climate to individual-level
Several researchers have found significant relationships servicescape–customer emotions is based on the idea that cus-
between service climate and positive customer perceptions of ser- tomers perceive how service venues present the servicescape’s
vice. A positive climate for employees’ service perceptions is linked substantive and communicative staging at the individual level.
to customer satisfaction (Schneider and Barbera, 2014). However, These perceptions then influence customer emotions. A good
neglecting the emotional components of satisfaction may not be service climate is an essential organizational environmental char-
sufficient for obtaining reliable predications of consumer responses acteristic, giving employees a reason why they should provide
(Martin et al., 2008; White and Yu, 2005). Although previous high-quality service. Therefore, service climate is an environ-
research focused on the service climate–customer satisfaction link mental characteristic variable at the firm level. As such, the
based on cognitive processes (e.g., beliefs about the target service), interaction between firm-level service climate and individual-level
White and Yu (2005) suggested an affective approach could be servicescape should substantially affect customer emotions. Thus,
used to examine service climate’s effect on customer satisfaction. two research hypotheses are:
When employees share common perceptions about their unit’s ser- H2-1. Firm-level service climate moderates the effect of perceived
vice quality (i.e., service climate), they should perform well with substantive staging of the servicescape on customer emotions at
customers, who will then report favorable employee performance the individual level, such that the effect is stronger when there is a
(Salanova et al., 2005). Although provision of excellent service positive service climate.
reflects a good service climate, there is little empirical research
investigating the relationship between service climate and cus- H2-2. Firm-level service climate moderates the effect of perceived
tomers’ emotional satisfaction in the hospitality field. Accordingly, communicative staging of the servicescape on customer emotions
the first hypothesis is: at the individual level, such that the effect is stronger when there
is a positive service climate.
H1. Firm-level service climate has a positive effect on customer
emotions at the individual level. 2.7.3. Cross-level direct effects: firm-level employee
engagement’s effect on individual-level behavioral intentions
2.7.2. Cross-level interaction effects: firm-level service climate’s Engagement is “the behavioral intensity and emotional qual-
effect on the relationship between individual-level servicescape ity of a person’s active involvement during a task” (Reeve et al.,
and customer emotions 2004, p. 147). Engaged employees have high energy and are enthu-
Arnould et al. (1998, p. 90) proposed that servicescapes are siastic about work (Bakker and Demerouti, 2008). Strong employee
produced by substantive and communicative staging. The for- motivation and high engagement with the organization is essential
mer are “physical creation of contrived environments,” and the when a service-industry firm pursues increased customer satis-
latter “ways in which the environment is presented and inter- faction and staff productivity, specifically in the service industry
preted” by service employees. According to Arnould et al. (1998), (Harter et al., 2002). Gonring (2008) claimed engaged employees
communicative staging is important for wilderness tourism (e.g., are more productive, profitable, customer-focused, and likely to
white water rafting), influencing customer experiences when the achieve customer satisfaction and loyalty, and less likely to exit.
servicescape is predominantly natural, not built. Such staging Research on employee engagement’s impact on customer
requires human intervention that increasingly evaluates, manages, behaviors has often focused on customer satisfaction. These stud-
and controls the wilderness. However, in theme parks, Dong and ies have inferred that employee engagement influenced customer
Siu (2013) found that the servicescape’s substantive staging (pri- loyalty and behavioral intentions, but rarely empirically explored
marily functional and mechanical) had a stronger influence on links between employee engagement and customer behavioral
customer service evaluations than communicative staging (human intentions. Discussions on the connection between employee
clues). These studies demonstrate evaluative polarization on the engagement and customer behaviors in practical reports, such as
nature–facilities continuum (Fredman et al., 2012), white water the Harvard Business Review Report (2013), used customer sat-
rafting’s natural environment attracts tourists focused on com- isfaction or corporate financial data to make inferences about
municative staging, while the opposite occurs in theme parks, the relationship between employee engagement and behavioral
which emphasize facilities. However, Lin and Mattila (2010) exam- intentions. Thus, although employee engagement is important
ined effects of both physical environment servicescape and service for keeping customers and increasing profits, there is little
encounters on customer emotions in restaurants, and suggested research examining effects of firm-level employee engagement on
120 K.-C. Chang / International Journal of Hospitality Management 53 (2016) 116–128

individual-level customer behavioral intentions in the hospitality Table 1

Hypothesis types and necessary conditions.
sector. Accordingly, the third hypothesis is:
Hypothesis types
H3. Firm-level employee engagement has a positive effect on cus-
1. Cross-level direct effect
tomer behavioral intentions at the individual level. Moderating variables at the firm level (service climate; employee
engagement) have direct effects on dependent variables at the individual level
2.7.4. Cross-level interaction effects: firm-level employee (H1: customer emotions; H3: behavioral intentions).
engagement’s effect on the relationship between individual-level 2. Cross-level moderator effect
Moderating variables at the firm level (service climate; employee
customer emotions and behavioral intentions engagement) moderate the relationships between variables at the individual
Based on the Mehrabian–Russell model (1974), which outlined level (H2: the relationship between substantive staging/communicative
the relationship between environmental stimuli and emotional staging of servicescape and customer emotions; H4: the relationship between
states, Lin and Liang (2011) summarized previous research, con- customer emotions and behavioral intentions).
cluding that service environments play an important role in service
Necessary conditions
delivery because they foster pleasant emotional reactions while
1. There is systematic within and between-group variance in customer
strengthening customer retention. White and Yu (2005) found emotions and behavioral intentions respectively.
positive emotions (e.g., happiness and pleasant surprises) were 2. There is significant variance at the level-1 intercept (customer emotions and
associated with increases in word-of-mouth referrals and will- behavioral intentions respectively).
3. There is significant variance in the level-1 slope (the effect of substantive
ingness to pay more, and negatively associated with switching
staging/communicative staging of servicescape on customer emotions; the
behavior. Because there are cognitive and affective components effect of customer emotions on behavioral intentions respectively).
of customer responses (White and Yu, 2005), Ladhari (2009) con- 4. For testing H1 and H3, the variance in the level-1 intercept is predicted by
cluded emotional satisfaction (e.g., happy and pleasant) was a service climate (for customer emotions) and employee engagement (for
better predictor of intentions (e.g., recommending a service) than behavioral intentions) respectively.
5. For testing H2 and H4, the variance in the level-1 slope is predicted by
satisfaction’s cognitive component (e.g., perceived service quality).
service climate (for the relationship between substantive
Although the customer emotions–behavioral intentions link is not staging/communicative staging of servicescape and customer emotions) and
new (Han and Back, 2007), this relationship serves as a founda- employee engagement (for the relationship between customer emotions and
tion for exploring effects of other variables on this relationship. For behavioral intentions) respectively.
example, Jani and Han (2015) tested the moderating effect of the
big five personality factors on customer emotion–loyalty relation-
3. Research method
ship.
Employee engagement means individuals are more deeply
3.1. Measurement
involved in their role (Kahn, 1990). The work engagement
expressed by one employee may fuel others’ engagement (Bakker
The items for each variable (listed in Table 2), measured on a 7-
et al., 2009), and engaged employees more successfully deliver ser-
point Likert scale from 1 = strongly disagree to 7 = strongly agree,
vice promises made by the firm (Zeithaml and Bitner, 2003).
were adapted for the context of a hot spring resort from prior
At the individual level, customers’ positive emotions originate
research, including studies on substantive/communicative stag-
from a good service environment (perceived through substan-
ing of the servicescape (Dong and Siu, 2013), customer emotions
tive staging of the servicescape) and/or good service quality
(Hennig-Thurau et al., 2006), customer behavioral intentions (Kuo
(perceived through communicative staging of the servicescape).
et al., 2012), service climate (Salanova et al., 2005), and employee
At the firm level, employees are energetic, focused, and dedi-
engagement (Britt et al., 2005). Sinaiko and Brislin’s (1973) back-
cated. Firms expect these qualities because engaged employees
translation procedure was used to ensure translation quality from
create satisfied and even delighted customers (Barnes et al.,
English to Chinese. Respondents also reported demographic char-
2014). The logic that relates firm-level employee engagement
acteristics (on categorical scales). Customers reported gender, age,
to individual-level customer emotions and behavior intentions
education, and visit frequency; Managerial employees gave gender,
is based on the notion that in addition to how customer emo-
age, education, and position; and employees reported gender, age,
tions at the individual level influence behavioral intentions toward
education, and work environment.
the firm, strong employee engagement is one of the reasons why
Initially, three experts (both academics and practitioners) in the
employees have a sense of responsibility and commitment to
hospitality field reviewed the instrument for content (i.e., face)
quality service. Thus, employee engagement can be a firm-level
validity (Hair et al., 2010), verifying items were related to the hot
characteristic, and the interaction between firm-level employee
spring resort context. Two items for the substantive staging of ser-
engagement and individual-level customer emotions should help
vicescape were deleted (i.e., layout makes facilities easy to use, and
determine behavioral intentions. Accordingly, the fourth hypothe-
signs are helpful) as not applicable to the context. After modifica-
sis is:
tions, the experts accepted the question content and wording. The
H4. Firm-level employee engagement moderates the effect of cus- three questionnaire types were pilot tested, each being adminis-
tomer emotions on behavioral intentions at the individual level, tered to 30 EMBA students in the hospitality department who had
such that the effect is stronger when there is positive employee been to hot spring resorts. After eliminating invalid questionnaires,
engagement. the remainder were retained for analysis. The survey instrument’s
scale reliability was deemed adequate (Cronbach’s alpha > 0.7)
2.7.5. Hypothesis types and necessary conditions (Hair et al., 2010), and the questionnaires were employed in the
Kozlowski and Klein (2000) indicated that variables with differ- current study.
ent levels of analysis in a single model result in nesting, causing
statistical bias. HLM is suited for analysis of nested data because it 3.2. Sampling target and procedure in the hot spring resort
accounts for both individual- and firm-level relationships, thereby context
identifying the relationship between predictor and outcome vari-
ables (Woltman et al., 2012). There are several necessary conditions The history of using hot springs has worldwide origins and dates
to support the hypotheses (Woltman et al., 2012), listed in Table 1, back to the earliest civilizations (LaMoreaux, 2005). Recently, hot
with results shown in Section 4.3 on hypothesis testing. and mineral spring-related services and activities earned more than
 K.-C. Chang / International Journal of Hospitality Management 53 (2016) 116–128 121

Table 2
Reliability and factor loadings of the measures.

Construct Factor loadings Composite Average variance

(>0.5) reliability (>0.7) extracted (%) (>50%)

Substantive staging of servicescape (ss)

ss1 The background music of the hot spring resort is pleasant. 0.692
ss2 The hot spring resort has nice smell. 0.822
ss3 The atmosphere of the hot spring resort is cheerful. 0.857
ss4 The hot spring resort is clean. 0.822
0.924 0.603
ss5 The hot spring resort has up-to-date facilities. 0.819
ss6 The architecture of the hot spring resort is attractive. 0.773
ss7 The color scheme of the hot spring resort is attractive. 0.660
ss8 The facilities of the hot spring resort are maintained well. 0.747

Communicative staging of servicescape (cs)

cs1 The employees of the hot spring resort are willing to help. 0.798
cs2 The employees of the hot spring resort are polite and friendly. 0.886
0.889 0.669
cs3 The employees of the hot spring resort give cusomers personal attention. 0.863
cs4 The employees of the hot spring resort are passionate. 0.714

Customer emotions (ce)

ce1 The hot spring resort makes me feel peppy. 0.750
ce2 The ho tspring resort makes me feel enthusiastic. 0.750 0.840 0.638
ce3 The hot spring resort makes me feel excited. 0.888

Behavioral intentions (bi)

bi1 I will return to the hot spring resort for my next trip. 0.823
bi2 I will recommend the hot spring resort to my family and friends. 0.882 0.862 0.676
bi3 I intend to stay in the hot spring resort within the next three years. 0.756

Service climate (sc)

sc1 Employee in our hot spring resort have knowledge of the job and the skills to 0.598
deliver superior work and service.
0.803 0.505
sc2 Employees in our hot spring resort receive the recognition and rewards for the 0.764
delivery of superior work and service.
sc3 The overall quality of service provided by our hot spring resort to customers is 0.714
excellent.
sc4 Employees in our hot spring resort are provided with tools, technology, and 0.712
other resources to support the delivery of quality work and service.

Employee engagement (ee)

ee1 I feel responsible for my job performance. 0.612
ee2 I am committed to my job. 0.797
0.801 0.504
ee3 How well I do in my job matters a great deal to me. 0.732
ee4 How I do in my job influences how I feel. 0.685

Note. All factor loadings are significant at: p < 0.001.

$50 billion worldwide in 2013 (Global Wellness Institute, 2014). data collection is to use a 20/30 ratio. This study thus selected
Taiwan possesses rich and varied hot spring resources because of 20 resorts at the firm level, and each selected 30 customers as
its specific geographic environment. This environment encourages individual-level respondents. Overall, 600 questionnaires were
construction of hot spring resorts and the development of facilities distributed to customers, 200 were distributed to managerial
that allow guests to enjoy the hot springs. To ensure a representa- employees, and 300 were distributed to employees.
tive sample, participants were recruited at hot spring resorts (see
Appendix A for examples) in Northern and Eastern Taiwan pro-
viding private lodging, public hot spring bathing with recreational 4. Data analysis and results
facilities, and restaurants. All included hot spring resorts received
a “hot spring certificate” that was issued from Taiwan Tourism Statistical analyses were conducted based on usable responses
Bureau. from the selected hot spring resorts (data were from 519 customers,
In addition, to avoid common method variance (Avolio et al., 177 managerial employees, and 270 employees). SPSS 18.0, AMOS
1991), data were collected over 3 months from customers (i.e., per- 18.0, and HLM 7.0 statistical software was used for empirical anal-
ceived servicescape, customer emotions, and behavioral intention), yses.
managerial employees (service climate), and employees (employee
engagement). The questionnaire packages with three categories
(i.e., customers, managerial employees, and employees) were sent 4.1. Descriptive analyses
to the selected resorts with the permission of management officers,
who helped to distribute the employee and managerial employee Of 519 customer questionnaires, 33.9% were from male and
questionnaires. Both groups of employees completed these during 65.7% were from female respondents. 82.3% respondents were
breaks. Front desk staff distributed the customer questionnaires to 25–44 years old, 73.4% held a bachelor’s degree, and 38% were ser-
guests, who anonymously completed them, returning them to the vice workers. A large of majority (75.1%) of respondents visited the
front desk. Respondents were thanked for participation and given location 2–4 times per year.
small gifts. To ensure sample representativeness, respondents were Of the 177 managerial employee questionnaires, 39% were from
limited to those aged 18 and older. male and 61% were from female respondents. 32.8% respondents
Bickle (2007) suggested that when a multilevel study includes were 35–44 years old, 59.9% held a bachelor’s degree, and the most
interactions between group and individual levels, one method of common occupation was a basic managerial position (81.9%).
122 K.-C. Chang / International Journal of Hospitality Management 53 (2016) 116–128

Table 3
Mean, standard deviation, and correlation coefficient of the constructs.

Construct Mean SD 1 2 3 4

Individual level
1. Substantive staging of servicescape 5.410 0.871 1.00
2. Communicative staging of servicescape 5.571 0.853 0.779** 1.00
3. Customer emotions 5.042 1.113 0.166** 0.169** 1.00
4. Behavioral intentions 5.134 1.076 0.195** 0.142** 0.768** 1.00

Firm level
1. Service climate 5.193 0.304 1.00
2. Employee engagement 5.329 0.260 0.323 1.00

Note. SD: standard deviation.

**
p < 0.01.

Of the 270 employee questionnaires, 37% were from male and For service climate, the measurement model showed that the fit
63% were from female respondents. 34.1% respondents were 25–34 statistics, like GFI (0.983), RMR (0.021), CFI (0.979), and CMIN/DF
years old, 60.4% held bachelor’s degrees, and half of the respondents (5.091/2 = 2.951), met suggested criteria for model goodness-of-fit.
(50.4%) reported they felt good about their work environment. For employee engagement, the measurement model showed GFI
(0.982), RMR (0.023), CFI (0.975), and CMIN/DF (9.346/2 = 4.673)
4.2. Reliability and validity testing met the suggested goodness-of-fit criteria. For service climate and
employee engagement, the model and observed data fit reasonably
The conceptual rationale for using an aggregate measure of well. Table 2 shows factor loadings, CR, and AVE values. Table 3
service climate and employee engagement at the firm level was reports means, standard deviations, and correlations for both anal-
discussed earlier. However, Kozlowski and Klein (2000) suggested ysis levels.
there should be statistical justification for aggregation. Bliese
(1998) argued that intraclass correlation coefficients (ICC1 and 4.3. Hypothesis testing
ICC2) and within-group interrater reliability coefficients (rwg ) can
justify aggregation to higher levels of analysis. ICCs assess the To assess the hypotheses, a sequence of models is required: null,
interrater reliability of respondents’ judgments. The ICC1 coef- random-coefficient regression, intercept-as-outcomes, and slopes-
ficient represents the proportion of variance in ratings at the as-outcomes (Wech and Heck, 2004). Also, Qin et al. (2014) suggest
individual level attributable to group membership. The ICC2 coef- it is difficult for researchers to obtain significant effects in empiri-
ficient represents the reliability of group-level means (Bliese, cal studies (e.g., Huang and Lee, 2012) when the group-level sample
1998). The results showed the ICC1 value was 0.461 for service size is relatively small. Thus, this study used traditional 0.05 signifi-
climate and 0.453 for employee engagement, exceeding the crite- cance levels to test relationships among variables at the same level,
rion of 0.12 (Bliese, 1998). The ICC2 value was 0.779 for service and 0.1 significance levels to test cross-level interaction and main
climate and 0.775 for employee engagement, meeting the 0.7 effects.
criterion (Bliese, 1998). The average rwg value was 0.886 for ser-
vice climate and 0.867 for employee engagement, meeting the
4.3.1. Null model
0.7 criterion (Bliese, 1998). Accordingly, these analyses justified
This study aims to explore effects of individual- and firm-
aggregating employee responses at the firm level for service cli-
level independent variables on customer emotions and behavioral
mate and employee engagement. Moreover, across all individual-
intentions; thus, HLM was used to examine cross-level effects
and firm-level variables, Cronbach’s alphas for internal consistency
between the two variable levels. First, a null model examined
reliability exceeded 0.7 (Hair et al., 2010), indicating they were
between-group variance in the dependent variables (i.e., customer
satisfactory.
emotions and behavioral intentions). Unless there is significant
CFA (confirmatory factor analysis) was conducted to assess
between-group variance in dependent variables, firm-level vari-
the convergent and construct validity of the measurement
ables (service climate and employee engagement) can explain
model for firm- and individual-level variables. Because cus-
significant amounts of such variance (Wech and Heck, 2004). Thus,
tomers and employees constitute different levels, sample sizes
the following null model without independent variables at level 1
differ. Thus, CFAs were separately conducted to examine valid-
or level 2 was estimated:
ity. For individual-level variables, CFA was performed with a
four-factor measurement model (perceived substantive staging of
servicescape, perceived communicative staging of servicescape, Level-1 model: Customer emotionsij = ˇ0j + rij
customer emotions, and behavioral intentions). In the initial anal- Level-2 model: ˇ0j =  00 + u0j
ysis, one item for customer emotions was deleted (the hot spring
resort makes me feel elated) because the factor loading was less As Table 4 shows, a significant chi-square for the dependent
than 0.4 (Hair et al., 2010). Also, because model fit was subopti- variable indicates the between-group variance ( 00 ) is significantly
mal, the measurement error modification index (MI) was used to different from zero (2 = 99.232, df = 19, p < 0.001), indicating that
delete items. Thus, two items on perceived substantive staging of the intercept term varies across groups (Wech and Heck, 2004).
servicescape (the physical environment pleases me; the style of The ICC1 was 0.123 (0.175/1.423), showing that 12.3% of the
decoration is fashionable) were excluded because their errors were between-group variance is in customer emotions. Thus, intraclass
highly correlated with other items, reflecting within-construct correlations were moderate (0.059–0.138), and multilevel analysis
error covariance (Hair et al., 2010). The final measurement model should be employed to further examine relations between vari-
showed that, except the GFI (0.877), the other statistics, like RMR ables (Cohen, 1988). The ICC2 (0.809 > 0.7) indicates the group can
(0.056), CFI (0.925), and CMIN/DF (624.876/129 = 4.844), met the be reliably differentiated based on group means (Bliese, 1998).
suggested model goodness-of-fit criteria. Thus, there was reason- Moreover, there was no significant change in the null model
able overall fit between the model and observed data. after controlling gender and age. Thus, it is appropriate to further
 K.-C. Chang / International Journal of Hospitality Management 53 (2016) 116–128 123

Table 4
HLM results of customer emotions as the dependent variable.

Construct Null model Control model Random model Intercept model Slope model

Individual level (1) SC CS SC CS SC CS

Intercept 5.045*** 5.042*** 5.047*** 5.047*** 5.047*** 5.047*** 5.047*** 5.047***
Substantive staging of servicescape (SC) 0.276** 0.278** 0.277*
Communicative staging of servicescape (CS) 0.273** 0.283** 0.274**

Firm level (2)

Service climate 0.741* 0.552† 0.851* 0.748*

Cross level
Service climate * Substantive staging 0.425*
Service climate * Communicative staging 0.709**

Control variable
Gender −0.040
Age 0.002
*
p < 0.05.
**
p < 0.01.
***
p < 0.001.
†
p < 0.1.

explore relationships between other predictors and the dependent only level-1 variables in the HLM model, with no level-2 predictors
variables for customer emotions. for ˇ0j and ˇ1j . A t-test determines significance of  10 . This provides
In addition, a null model was estimated for the dependent evidence for whether the pooled level-1 slopes between indepen-
variable (behavioral intentions), with no independent variables at dent and dependent variables are non-zero. A random-coefficient
levels 1 or 2 as follows: regression model was estimated with the following:

Level-1 model: Behavioral intentionsij = ˇ0j + rij Level-1 model: Customer emotionsij = ˇ0j + ˇ1j *(Substantive staging
Level-2 model: ˇ0j =  00 + u0j of servicescapeij ) + rij
Level-2 model: ˇ0j =  00 + u0j ˇ1j =  10 + u1j
Table 5 results show a significant chi-square for the dependent
variable, indicating that between-group variance ( 00 ) is signifi- As Table 4 shows, the value for substantive staging of the ser-
cantly different from zero (2 = 81.184, df = 19, p < 0.001) and that vicescape ( 10 ) is 0.276 (t = 3.260, p < 0.01), meaning this variable
the intercept term varies across groups (Wech and Heck, 2004). has a positive significant effect on customer emotions. More-
The ICC1 is 0.111 (0.129/1.162 = 0.111), indicating that 11% of over, the chi-square for estimating the intercept ( 00 ) is 0.179
between-group variance in behavioral intentions is not explained (2 = 112.399, df = 19, p < 0.001), indicating there is significant vari-
and requires further examination using HLM (Cohen, 1988). The ance in intercepts and slopes of the dependent variable across
ICC2 (0.764 > 0.7) represents acceptable reliability for the group groups.
means (Bliese, 1998). There was no significant change in the null In addition, to examine the relationship between commu-
model after controlling for gender and age. Thus, it is appropriate nicative staging of the servicescape and customer emotions, a
to explore further the relationships between other predictors and random-coefficient regression model was estimated with the fol-
the dependent variables for behavioral intentions. lowing:

4.3.2. Random-coefficient regression model Level-1 model: Customer emotionsij = ˇ0j + ˇ1j *(Communicative
As discussed above, to verify the perceived staging of servicescapeij ) + rij
servicescape–customer emotions–behavioral intentions rela- Level-2 model: ˇ0j =  00 + u0j ˇ1j =  10
tionship, it is important to examine cross-level effects. Thus,
confirming the S–O–R relationship requires significant between- As Table 4 shows, the value of communicative staging of the
group variation in intercepts and slopes using a random-coefficient servicescape ( 10 ) is 0.273 (t = 3.095, df = 19, p < 0.01), showing this
regression model (Wech and Heck, 2004). Such models include variable significantly affects customer emotions. The chi-square

Table 5
HLM results of behavioral intentions as the dependent variable.

Construct Null model Control model Random model Intercept model Slope model

Individual level (1)

Intercept 5.142*** 5.140*** 5.142*** 5.142*** 5.142***
Customer emotions 0.733*** 0.726*** 0.731***

Firm level (2)

Employee engagement 0.874*** 0.770***

Cross level
Customer emotions * Employee engagement 0.220†

Control variable
Gender 0.140
Age 0.107
***
p < 0.001.
†
p < 0.1.
124 K.-C. Chang / International Journal of Hospitality Management 53 (2016) 116–128

for estimating the intercept ( 00 ) is 0.178 (2 = 112.744, df = 19, moderates the relationship between substantive staging of the ser-
p < 0.001), confirming there is significant variance in intercepts and vicescape and customer emotions. Moreover, the deviance of the
slopes of the dependent variable across groups. null model decreased from 1544.007 to 1487.777 in the slope-as-
Furthermore, to examine the customer emotions–behavioral outcomes model, indicating this model had better fit to the data,
intentions link, a random-coefficient regression model was esti- supporting Hypothesis 2-1.
mated with the following: In addition, there is a significant cross-level interaction between
service climate and communicative staging of the servicescape
Level-1 model: Behavioral intentionsij = ˇ0j + ˇ1j *(Customer ( 11 = 0.709, t = 2.964, p < 0.01), indicating service climate mod-
emotionsij ) + rij erates the relationship between communicative staging of the
Level-2 model: ˇ0j =  00 + u0j ˇ1j =  10 + u1j servicescape and customer emotions. Moreover, the deviance of
the null model decreased from 1544.007 to 1481.541 in the slope-
As Table 5 shows, the value of customer emotions ( 10 ) is 0.733 as-outcome model, indicating better model fit, and supporting
(t = 19.598, df = 19, p < 0.001), indicating customer emotions sig- Hypothesis 2-2.
nificantly affect behavioral intentions. Moreover, the chi-square For relationships among employee engagement, customer emo-
for estimating the intercept ( 00 ) is 0.151 (2 = 182.662, df = 19, tions, and behavioral intentions, the slope-as-outcomes model was
p < 0.001), confirming there is significant variance in the intercepts estimated using:
and slopes for the dependent variable across groups.
Level 1: Behavioral intentionsij = ˇ0j + ˇ1j *(Customer emotionsij ) + rij
4.3.3. Intercept-as-outcomes model Level 2: ˇ0j =  00 +  01 *(Employee engagementj ) + u0j
After establishing significant variation across groups for level-1 ˇ1j =  10 +  11 *(Employee engagementj ) + u1j
intercepts, cross-level Hypotheses 1 and 3 can be tested as follows:
Table 5 shows a significant cross-level interaction between
Level-1 model: Customer emotionsij = ˇ0j + ˇ1j *(Substantive staging employee engagement and customer emotions ( 11 = 0.220,
of servicescapeij ) + rij t = 1.916, p < 0.1), suggesting employee engagement moderates the
Level-2 model: ˇ0j =  00 +  01 *(Service climatej ) + u0j ˇ1j =  10 customer emotions–behavioral intentions relationship, supporting
Level-1 model: Customer emotionsij = ˇ0j + ˇ1j *(Communicative Hypothesis 4. Moreover, the deviance of the null model decreased
staging of servicescapeij ) + rij from 1520.784 to 1120.662 in the slope-as-outcome model, indi-
Level-2 model: ˇ0j =  00 +  01 *(Service climatej ) + u0j ˇ1j =  10 cating better fit for the latter model.

As Table 4 shows, for communicative staging of the servicescape,

5. Conclusions and implications
the service climate value ( 01 ) is 0.741 (t = 2.148, df = 18, p < 0.05),
indicating service climate significantly affects customer emotions.
5.1. Conclusions
Moreover, for the communicative staging of the servicescape,
the service climate value ( 01 ) is 0.552 (t = 2.090, df = 18, p < 0.1),
To the best of our knowledge, few hospitality studies show
indicating service climate also significantly impacts customer emo-
how firm-level moderating effects (service climate and employee
tions, supporting Hypothesis 1.
engagement) affect relationships among perceived servicescape,
The below equations test customer emotions. Table 5 shows
customer emotions, and behavioral intentions at the individual
the employee engagement value ( 01 ) is 0.874 (t = 4.133, df = 18,
level. This study’s results indicate that substantive staging of the
p < 0.001), indicating employee engagement significantly affects
servicescape of hot spring resorts, as perceived by customers,
behavioral intentions, supporting Hypothesis 3.
positively affects customer emotions ( 10 = 0.276, p < 0.01). Fur-
thermore, perceived communicative staging of the servicescape
Level-1 model: Behavioral intentionsij = ˇ0j + ˇ1j *(Customer engenders an equivalent effect on customer emotions ( 10 = 0.273,
emotionsij ) + rij p < 0.01), suggesting employee behavior cannot be ignored even
Level-2 model: ˇ0j =  00 +  01 *(Employee engagementj ) + u0j when firms provide customers with luxurious tangibles. The results
ˇ1j =  10 imply that tangible facilities and intangible services within such
resorts interact, creating value for customers, consistent with
4.3.4. Slope-as-outcomes model the notion “marketing intangible products and product intangi-
Finally, after establishing significant group variance in slopes bles” (Levitt, 1981). Although hotel businesses carry equal weight
using random-coefficient regression models, it is important to in tangibles and intangibles (Chang and Tarn, 2008), service’s
examine whether variance in the slope across groups is sig- intangibility provides opportunities for the hospitality service
nificantly related to group-level independent variables. The providers to create unique experiences for guests (Ford et al.,
slope-as-outcomes model was estimated by: 2012). Consequently, as Mehrabian and Russell (1974) noted, cus-
tomers’ emotional satisfaction affects their behavioral intentions
Level 1: Customer emotionsij = ˇ0j + ˇ1j *(Substantive staging of ( 10 = 0.733, p < 0.01) toward hot spring resorts.
servicescapeij ) + rij Second, the results suggest that firm-level service climate
Level 2: ˇ0j =  00 +  01 *(Service climatej ) + u0j positively affects individual-level customer emotions for the sub-
ˇ1j =  10 +  11 *(Service climatej ) + u1j stantive ( 01 = 0.741, p < 0.05) and the communicative ( 01 = 0.552,
Level 1: Customer emotionsij = ˇ0j + ˇ1j *(Communicative staging of p < 0.1; H1 was supported) staging of the servicescape. Ser-
servicescapeij ) + rij vice climate has a stronger direct effect on substantive than
Level 2: ˇ0j =  00 +  01 *(Service climatej ) + u0j on communicative staging of the servicescape. H2-2 was sup-
ˇ1j =  10 +  11 *(Service climatej ) + u1j ported by a cross-level interaction between service climate and
communicative staging of the servicescape, evidencing stronger
As Table 4 shows, there is significant cross-level interaction positive effects on customer emotions ( 11 = 0.709, p < 0.01) com-
between service climate and substantive staging of the ser- pared to the cross-level interaction between service climate and
vicescape (␥11 = 0.425, t = 2.305, p < 0.05), indicating service climate perceived substantive staging of the servicescape ( 11 = 0.425,
 K.-C. Chang / International Journal of Hospitality Management 53 (2016) 116–128 125

p < 0.05; supporting H2-1). This is consistent with Kralj and Solnet the work group where subjects are nested. Therefore, the present
(2010)’s finding of high correlations between internal organiza- study elucidated the contribution of employee behaviors in ser-
tional dynamics (service climate) and customer satisfaction (an vice delivery by clarifying the why (service climate and employee
important business outcome). In conclusion, our results for service engagement at firm level) and how (perceived servicescape and
climate are consistent with the service marketing triangle model, customer emotions at individual level) of their interactions with
which states service firms must provide employees with neces- customers. This clarification is critical for understanding the influ-
sary skills, resources, and rewards to demonstrate the service firm’s ence of customer emotions and behavioral intentions toward a
commitment to service quality – a condition that, in turn, leads to service firm.
better customer service (Bitner, 1995).
Also noteworthy is that at hot spring resorts, firm-level 5.3. Managerial implications
employee engagement positively affects individual-level customer
emotions ( 01 = 0.874, p < 0.01; supporting H3). It also indicates Perceptions of the servicescape (e.g., music, odor, color, equip-
that cross-level interaction between employee engagement and ment, and architecture) may help customers distinguish and
customer emotions ( 11 = 0.220, p < 0.1; supporting H4) positively categorize firms; servicescapes can be indicators of expected qual-
affects behavioral intentions. In emotional contagion theory, the ity (Bitner, 1992). Many businesses (e.g., shopping centers and
ripple effect suggests that engaged employees who reach personal nightclubs) recognize that the service’s physical environment
or company goals elicit good moods in customers (Bakker et al., effects customer experience, and have thus developed the envi-
2009; Barsade, 2002). Feedback theory proposes that employee ronment (Zomerdijk and Voss, 2010). Thus, the substantive staging
perceptions of customer delight increase employee engagement of the servicescape, it should be compatible with guests’ expec-
(Barnes et al., 2014). Thus, cross-level interactions occur between tations (Lin, 2004). Every service provider must develop a unique
employee engagement and customer emotions, leading to positive servicescape best representing and visualizing its intangible com-
behavioral intentions. petencies and qualities (Reimer and Kuehn, 2005).
Kaltcheva and Weitz (2006) proposed that service providers
5.2. Theoretical implications could use different elements (e.g., music, color, and texture) in their
store to influence customers’ arousal levels. These elements could
This study’s results theoretically contribute to the existing ser- be stimuli engendering customer perceptions, like affective qual-
vicescape literature. First, research has found different results for ity perceptions of pleasure-displeasure and activating-deactivating
how substantive and communicative stagings of the servicescape (Russell, 2003). Affect is an “emotion expressed in language and
influence customer experiences (e.g., Arnould et al., 1998; Dong the affective quality of an environment is the emotion-inducing
and Siu, 2013). However, the present results are similar to Lin quality that persons verbally attribute to that place” (Russell and
and Mattila (2010), in which servicescapes of the physical envi- Pratt, 1980, p. 311–312). To foster enjoyable customer experiences,
ronment and service encounters with employees had equivalent hot spring resorts could choose different design themes like ori-
effects on customer pleasure. Therefore, hospitality firms should ental, modern simplicity, old times or tropical, to “tell stories”
note that consumer responses to servicescape stimuli (e.g., physi- (Rosenbaum and Massiah, 2011). These should be consistent in
cal cues and service encounters) represent what actually happens color, music, and decoration to help identify a resort’s unique-
from customers’ point of view (Fredman et al., 2012; Zomerdijk and ness. The servicescape—through architecture, lobby, rooms, hot
Voss, 2010). spring bathing or amusement facilities—can elicit customers’ affec-
Second, Mari and Poggesi (2013) reviewed 188 servicescape- tive responses, e.g., laid-back, romantic, nostalgic, or exotic feelings
related studies and suggested future research should continue that make the customers experience “love at first sight” (Zhang and
exploring this topic, with emphasis on employee motivation, Li, 2004). This response is an important starting point for creating
satisfaction, productivity, and their influence on customer behav- positive emotions for customers entering hot spring resorts.
ior. In response, this study used the service marketing triangle Furthermore, the communicative staging of a servicescape
framework (Bitner, 1995), including firm-level variables of ser- has an increasingly important function that, with substantive
vice climate and employee engagement, to empirically clarify their staging, influences positive emotions, engendering positive behav-
effects on individual-level relations with servicescape, customer ioral intentions. Customers visit hot spring resorts for diverse
emotions, and behavioral intentions (Bitner, 1992) based on the leisure activities; thus, service encounters between customers and
Mehrabian–Russell model (1974). employees are frequent; customers, during hot spring activities
Finally, in HLM analyses, it is possible for researchers to perceive like dining, requesting towels, adjusting water temperature or even
that individuals are subject to clustering effects at higher levels (like shopping, are influenced by service provided by staff. Resort man-
the organization); however, single-level analyses ignoring nesting agers should recognize that customer experiences are synthesized
can cause erroneous estimates (e.g., Type I errors, indicating overly when employees deliver services complementing the physical
lax statistical decisions, or Type II errors, indicating overly strin- environment.
gent statistical decisions) and inappropriate conclusions (Bliese and However, hospitality industry employees face inherent chal-
Hanges, 2004; Lin and Peng, 2006). For example, past research has lenges (low pay, long hours, and turnover) (Chiang and Jang,
used single-level models to examine how service climate connects 2008), possibly influencing work motivation, attitudes, and behav-
to customer outcomes (e.g., Mokhtaran et al., 2015), but this ignores ior. Thus, creating a good service climate and motivating employee
potentially meaningful information at the firm and individual lev- engagement are key to employee performance, as is training for
els (Wang, 2015). The HLM design avoids the ecological fallacy of employees (e.g., programs increasing their understanding of the
making inferences about smaller units of analysis based on the water qualities and curative effects of different types of springs,
characteristics of larger units of analysis, while also avoiding the massage, facial and nail beauty skills, regimens and health benefits,
individualistic fallacy of making inferences about larger units of and foreign languages), making them feel supported by manage-
analysis based on the characteristics of smaller units of analysis ment so they can satisfy customer requirements. Hot spring resort
(Lin and Peng, 2006; Robinson, 1950). managers must provide rewards and recognition encouraging
Wieseke et al. (2008) deemed multilevel methodology appro- employees to reciprocate with greater engagement (Saks, 2006).
priate for studies examining how organizational-level variables Therefore, this study has implications for hot spring resort opera-
explain individual-level behavior. The former likely depend on tors, as it indicates why employees motivate themselves to work, a
126 K.-C. Chang / International Journal of Hospitality Management 53 (2016) 116–128

prerequisite for determining how they can serve customers well, customers dislike the sulfur scent but tolerate it for its curative and
engendering positive customer emotions and behavioral inten- rejuvenating power. Therefore, further investigation is needed into
tions. how different hot spring types influences customer perceptions
through scent.
6. Limitations and future research directions Furthermore, because nature-based servicescapes (Fredman
et al., 2012) are not assessed by the current substantive staging
Although the measurement scale for the substantive staging of the servicescape measure, items assessing hot spring resorts’
of servicescape was modified from previous research, there were natural surroundings (e.g., the surrounding natural landscapes is
challenges in applying it to examine hot spring resorts. Thus, the beautiful) should be examined for influence on customers’ percep-
coefficient of variation (standard deviation/mean) for each variable tions of the servicescape in this context; those inclined toward
(i.e., music, smell, atmosphere, cleanliness, up-to-date facilities, natural feelings and those preferring constructed facilities may
architecture, color scheme, and well-maintained facilities) was cal- have different opinions when evaluating hot spring resort ser-
culated to understand the extent of dispersion in the data. The vicescapes.
results show dispersion is more prominent for smell; because hot
and mineral springs are the main attractions in such resorts, the Acknowledgments
measurement for “having a nice smell in the hot spring resort”
may have been influenced by different types of springs. For exam- The author would like to thank anonymous reviewers for useful
ple, although sodium bicarbonate hot springs, being colorless and suggestions and the Ministry of Science and Technology of Taiwan
odorless, do not influence ambient scent, sulfur springs do. Some (R.O.C.) for financial support (MOST 104-2410-H-263-007).

Appendix A.
 K.-C. Chang / International Journal of Hospitality Management 53 (2016) 116–128 127

References Karantinou, K.M., Hogg, M.K., 2001. Exploring relationship management in

professional service: a study of management consultancy. J. Mark. Manage. 17
Arnould, E.J., Price, L.A., Tierney, P., 1998. Communicative staging of the wilderness (3/4), 263–286.
servicescape. Serv. Ind. J. 18 (3), 90–115. Kozlowski, S.W.J., Klein, K.J., 2000. A multilevel approach to theory and research in
Avolio, B.J., Yammarino, F.J., Bass, B.M., 1991. Identifying common methods organizations: contextual, temporal, and emergent processes. In: Klein, K.J.,
variance with data collected from a single source: an unresolved sticky issue. J. Kozlowski, S.W.J. (Eds.), Multilevel Theory, Research and Methods in
Manage. 17 (3), 571–587. Organizations: Foundations, Extensions, and New Directions. San Francisco,
Bagozzi, R.P., 1975. Marketing as exchange. J. Mark. 39 (4), 32–39. CA, Jossey-Bass, pp. 3–90.
Bagozzi, R.P., Gopinath, M., Nyer, P.U., 1999. The role of emotions in marketing. J. Kralj, A., Solnet, D., 2010. Service climate and customer satisfaction in a casino
Acad. Mark. Sci. 27 (2), 184–206. hotel: an exploratory case study. Int. J. Hosp. Manage. 29 (4), 711–719.
Bakker, A.B., Demerouti, E., 2008. Towards a model of work engagement. Career Kuo, N.T., Chang, K.C., Chen, M.C., Hsu, C.L., 2012. Investigating the effect of service
Dev. Int. 13 (3), 209–223. quality on customer post-purchasing behaviors in the hotel sector: the
Bakker, A.B., Westman, M., van Emmerik, I.J.H., 2009. Advancements in crossover moderating role of service convenience. J. Qual. Assur. Hosp. Tour. 13 (3),
theory. J. Manage. Psychol. 24 (3), 206–219. 212–234.
Barnes, D.C., Collier, J.E., Robinson, S., 2014. Customer delight and work Ladhari, R., 2009. Service quality, emotional satisfaction, and behavioral intentions.
engagement. J. Serv. Mark. 28 (5), 380–390. Manag. Serv. Qual. 19 (3), 308–331.
Barsade, S., 2002. The ripple effect: emotional contagion and its influence on group LaMoreaux, P.E., 2005. History and classification of springs. Geol. Soc. Am. Abstr.
behavior. Adm. Sci. Q. 47 (4), 644–675. Prog. 37 (7), 324.
Bickle, R., 2007. Multilevel Analysis for Applied Research: It’s Just Regression. Levitt, T., 1981. Marketing intangible products and product intangibles. Harv. Bus.
Guilford Press, New York. Rev. 59 (3), 94–102.
Bitner, M.J., 1992. Servicescape: the impact of physical surroundings on customers Lewin, K., 1951. Field Theory in Social Science. Harper & Row, New York.
and employees. J. Mark. 56 (2), 57–71. Liao, H., Chuang, A., 2004. A multilevel investigation of factors influencing
Bitner, M.J., 1995. Building service relationships: it’s all about promises. J. Acad. employee service performance and customer outcomes. Acad. Manage. J. 47
Mark. Sci. 23 (4), 246–251. (1), 41–58.
Bliese, P.D., 1998. Group size, ICC values, and group-level correlations: a Lin, I.Y., 2004. Evaluating a serviscape: the effect of cognition and emotion. Int. J.
simulation. Org. Res. Methods 1 (4), 355–373. Hosp. Manage. 23 (2), 163–178.
Bliese, P.D., Hanges, P.J., 2004. Being both too liberal and too conservative: the Lin, J.S.C., Liang, H.Y., 2011. The influence of service environments on customer
perils of treating grouped data as though they were independent. Org. Res. emotion and service outcomes. Manag. Serv. Qual. 21 (4), 350–372.
Methods 7 (4), 400–417. Lin, I.Y., Mattila, A.S., 2010. Restaurant servicescape, service encounter, and
Britt, T.W., 1999. Engaging the self in the field: testing the triangle model of perceived congruency on customers’ emotions and satisfaction. J. Hosp. Mark.
responsibility. Pers. Soc. Psychol. Bull. 25 (6), 696–706. Manage. 19 (8), 819–841.
Britt, T.W., Castro, C.A., Adler, A.B., 2005. Self-engagement, stressors, and health: a Lin, C.C., Peng, T.K., 2006. Multilevel research in management: conceptual
longitudinal study. Pers. Soc. Psychol. Bull. 31 (11), 1475–1486. theoretical, and methodological issues in level of analysis. J. Manage. 23 (6),
Britt, T.W., Dickinson, J.M., Greene-Shortridge, T.M., McKibben, E.S., 2007. 649–675.
Self-engagement at work. In: Nelson, D., Cooper, C. (Eds.), Positive Lo, A.S., Wu, C., 2014. Effect of consumption emotion on hotel and resort spa
Organizational Behavior. SAGE Publications, London, pp. 143–159. experience. J. Travel Tour. Mark. 31 (8), 958–984.
Cha, J., Borchgrevink, C.P., 2014. Service climate in restaurants. Int. J. Hosp. Tour. Manning, M.L., Davidson, M., Manning, R.L., 2005. Measuring tourism and
Adm. 15 (1), 19–37. hospitality employee workplace perceptions. Int. J. Hosp. Manage. 24 (1),
Chang, K.C., Tarn, D.D.C., 2008. Does service tangibilization work in the hotel 75–90.
industry? An experimental investigation. Asia Pac. J. Tour. Res. 13 (4), 411–434. Mari, M., Poggesi, S., 2013. Servicescape cues and customer behavior: a systematic
Chiang, C.F., Jang, S.C., 2008. An expectancy theory model for hotel employee literature review and research agenda. Serv. Ind. J. 33 (2), 171–199.
motivation. Int. J. Hosp. Manage. 27 (2), 313–322. Martin, D., O’Neill, M., Hubbard, S., Palmer, A., 2008. The role of emotion in
Cohen, J., 1988. Statistical Power Analysis for the Behavioral Sciences, 2nd ed. explaining consumer satisfaction and future behavioral intention. J. Serv. Mark.
Eribaum, Hillsdale, NJ. 22 (3), 224–236.
Dong, P., Siu, N.Y.M., 2013. Servicescape elements, customer predispositions and May, D.R., Gilson, R.L., Harter, L.M., 2004. The psychological conditions of
service experience: the case of theme park visitors. Tour. Manage. 36, meaningfulness, safety and availability and the engagement of the human
541–551. spirit at work. J. Occup. Psychol. 77 (1), 11–37.
Donovan, R., Rossiter, J.R., 1982. Store atmosphere: an environmental psychology Mehrabian, A., Russell, A., 1974. An Approach to Environmental Psychology. MIT
approach. J. Retail. 58 (1), 34–57. Press, Cambridge, MA.
Fishbein, M., Ajzen, I., 1975. Belief, Attitude, Intention and Behavior: An Mokhtaran, M., Fakharyan, M., Jalilvand, M.R., Mohebi, M., 2015. The effect of
Introduction to Theory and Research. Addison-Weskey, Reading, MA. service climate on perceived service value and behavioral intentions: the
Ford, R.C., Sturman, M.C., Heaton, C.P., 2012. Managing Quality Service in mediating role of service quality. Asia Pac. J. Tour. Res. 20 (4), 472–486.
Hospitality: How Organizations Achieve Excellence in the Guest Experience. Nilsson, E., Ballantyne, D., 2014. Reexamining the place of servicescape in
Cengage Learning, Clifton Park, NY. marketing: a service-dominant logic perspective. J. Serv. Mark. 28 (5), 374–379.
Fredman, P., Wall-Reinius, S., Grundén, A., 2012. The nature of nature in Qin, Q., Wen, B., Ling, Q., Zhou, S., Tong, M., 2014. How and when the effect of
nature-based tourism. Scand. J. Hosp. Tour. 12 (4), 289–309. ethical leadership occurs? A multilevel analysis in the Chinese hospitality
Global Wellness Institute, 2014. Global Spa & Wellness Economy Monitor, http:// industry. Int. J. Contemp. Hosp. Manage. 26 (6), 974–1001.
mrot.pl/images/pliki/GWI Global Spa and Wellness Economy Monitor Full Reeve, J., Jang, H., Carrell, D., Jeon, S., Barch, J., 2004. Enhancing students’
Report .pdf (accessed July 28). engagement by increasing teachers’ autonomy support. Motiv. Emot. 28 (2),
Gonring, M.P., 2008. Customer loyalty and employee engagement: an alignment 147–169.
for value. J. Bus. Strategy 29 (4), 29–40. Reimer, A., Kuehn, R., 2005. The impact of servicescape on quality perception. Eur.
Hair, J.F., Black, W.C., Babin, B.J., Anderson, R.E., 2010. Multivariate Data Analysis: A J. Mark. 39 (7/8), 785–808.
Global Perspective. Prentice Hall, Upper Sadder River, NJ. Robinson, W.S., 1950. Ecological correlations and the behavior of individuals. Am.
Han, H., Back, K.J., 2007. Assessing customers’ emotional experiences influencing Soc. Rev. 15 (3), 351–357.
their satisfaction in the lodging industry. J. Travel Tour. Mark. 23 (1), 43–56. Rosenbaum, M.S., Massiah, C., 2011. An expanded servicescape perspective. J. Serv.
Harter, J.K., Schmidt, F.L., Hayes, T.L., 2002. Business-unit-level relationship Manage. 22 (4), 471–490.
between employee satisfaction, employee engagement, and business Russell, J.A., 2003. Core affect and the psychological construction of emotion.
outcomes: a meta-analysis. J. Appl. Psychol. 87 (2), 268–279. Psychol. Rev. 110 (1), 145–172.
Harvard Business Review Report, 2013. The Impact of Employee Engagement on Russell, J.A., Pratt, G., 1980. A description of the affective quality attributed to
Performance. Harvard Business School Publishing, Boston, MA. envirnments. J. Pers. Soc. Psychol. 38 (2), 311–322.
Hennig-Thurau, T., Groth, M., Paul, M., Gremler, D.D., 2006. Are all smiles created Saks, A.M., 2006. Antecedents and consequents of employee engagement. J.
equal? How emotional contagion and emotional labor affect service Manage. Psychol. 21 (7), 600–619.
relationships. J. Mark. 70 (3), 57–73. Salanova, M., Agut, S., Peiró, J.M., 2005. Linking organizational resources and work
Heskett, J.L., Jones, T.O., Loveman, G.W., Saser, W.E., Schlesinger, L.A., 1994. Putting engagement to employee performance and customer loyalty: the mediation of
the service-profit-chain to work. Harv. Bus. Rev. 72 (2), 164–174. service climate. J. Appl. Psychol. 90 (6), 217–227.
Huang, M.C., Lee, P.P., 2012. The moderating effects of internal marketing and Schneider, B., Barbera, K., 2014. The Oxford Handbook of Organizational Climate
external markeitng upon the relationship between interactive marketing and and Culture. Oxford University Press, New York.
customer satisfaction: a multi-level model. Sun Yat-Sen Manage. Rev. 20 (2), Schneider, B., Gunnarson, S.K., Niles-Jolly, K., 1994. Creating the climate and
637–672. culture of success. Org. Dyn. 23 (1), 17–29.
Jani, D., Han, H., 2015. Influence of environmental stimuli on hotel customer Schneider, B., Parkington, J., Buxton, V., 1980. Employee and customer perceptions
emotional loyalty response: testing the moderating effect of the big five of service in banks. Adm. Sci. Q. 25 (2), 252–267.
personality factors. Int. J. Hosp. Manage. 44 (January), 48–57. Schneider, B., Salvaggio, A.N., Subirats, M., 2002. Climate strength: a new direction
Kahn, W.A., 1990. Psychological conditions of personal engagement and for climate research. J. Appl. Psychol. 87 (2), 220–229.
disengagement at work. Acad. Manage. J. 33 (4), 692–724. Schneider, B., White, S.S., Paul, M.C., 1998. Linking service climate and customer
Kaltcheva, V.D., Weitz, B.A., 2006. When should a retailer create an exciting store perceptions of service quality: test of a causal model. J. Appl. Psychol. 83 (2),
environment? J. Mark. 70 (1), 107–118. 150–163.
128 K.-C. Chang / International Journal of Hospitality Management 53 (2016) 116–128

Sinaiko, H.W., Brislin, R.W., 1973. Evaluating language translations: experiments Wieseke, J., Lee, N., Broderick, A.J., Dawson, J.F., Dick, R.V., 2008. Multilevel analysis
on three assessment methods. J. Appl. Psychol. 57 (3), 328–334. in marketing research: differentiating analytical outcome. J. Mark. Theory
Slåtten, T., Mehmetoglu, M., 2011. Antecedents and effects of engaged frontline Pract. 16 (4), 321–339.
employees: a study from the hospitality industry. Manag. Serv. Qual. 21 (1), Woltman, H., Feldstain, A., MacKay, J.C., Rocchi, M., 2012. An introduction to
88–107. hierarchical linear modeling. Tutor. Quant. Methods Psychol. 8 (1), 52–69.
Valentine, P., 1992. Review: nature-based tourism. In: Weiler, B., Hall, C.M. (Eds.), Zeithaml, V.A., Berry, L.L., Parasuraman, A., 1996. The behavioral consequences of
Special Interest Tourism. Belhaven Press, London, Great Britain, pp. 105–127. service quality. J. Mark. 60 (2), 31–46.
Wakefield, K.L., Blodgett, J.G., 1994. The importance of servicescapes in leisure Zeithaml, V.A., Bitner, M.J., 2003. Service Marketing: Integrating Customer Focus
service settings. J. Serv. Mark. 8 (3), 66–76. across the Firm. McGraw-Hill, New York.
Wang, M.L., 2015. Linking service climate to customer loyalty. Serv. Ind. J. 35 (7/8), Zhang, P., Li, N., 2004. Love at first sight or sustained effect? The role of perceived
403–414. affective quality on users’cognitive reactions of information technology. In:
Warshaw, P.R., Davis, F.D., 1985. Disentangling behavioral intentions and Paper presented at the Proceedings of 25th International Conference on
behavioral expectations. J. Exp. Soc. Psychol. 21 (3), 213–228. Information Systems, Washington, DC.
Wech, B.A., Heck, A.L., 2004. An introduction to hierarchical linear modeling for Zomerdijk, L.G., Voss, C.A., 2010. Service design for experience-centric services. J.
marketing researchers. Mark. Bull. 15, 1–8, Technical Note 1. Serv. Res. 13 (1), 67–82.
White, C., Yu, Y., 2005. Satisfaction emotions and consumer behavioral intentions.
J. Serv. Mark. 19 (6), 411–420.