Dance with Problem Customers: An Examination of the Moderating Effect of Emotional Intelligence

Chiao Da Management Review Vol. 33 No.1. 2013 pp.1-36

Dance with Problem Customers: An Examination of the

m宜。derating

Effect of Emotional Intelligence

與奧客共舞:情緒智力調節效果檢驗

胡秀華 1 Hsiu-Hua Hu

銘傳大學國際企業學系

Department of Intemational Business, Ming Chuan University.

摘要:本研究的目的條透過資源保存理論的觀點，來探討一線員工的問題顧客 知覺、情緒智力、情緒勞務負擔、與其正向服務行為之間的關聯性。本研究 以台灣八十家中型餐廳、之一線員工及其直屬主管為研究對象，共回收232份配 對問卷，經統計分析後，得到以下結論:一線員工的問題顧客知覺會正向影 響其情緒勞務負擔，並負向影響其正向服務行為。而且，一線員工的情緒勞 務負擔在問題顧客知覺與正向服務行為問扮演部份中介角色。同時，一線員 工的高情緒智力可分別調節問題顧客知覺與情緒勞動負擔間的正向關條、以 及問題顧客知覺與正向服務行為間的負向關條。最後，本研究根據分析結果 提出實務意涵、研究限制、與未來研究的建議。 關鍵字:資源保存坦言侖;情緒智力;情緒勞務負擔;正向服務行為;問題顧客知覺

Abstract: Adopting a perspective of conservation of resources, the aim of this study is to examine the association among 企ontline employees' problem customer perceptions, emotional intelligence, emotional labor tax, and positive service behaviors. This study uses data collected from a field survey of 232 employee-supervisor matched dyads from 80 mid-size restaurants in Taiwan to test a 企amework of influences of problem customer perceptions. The results indicate that problem customer perceptions are positively related to 企ontline

1 Corresponding author: Department of International Business, Ming Chuan Universi旬， Taipei City, Taiwan, E-mail: shhu@mail.mcu.edu.tw

01 the Moderating ξffect 01 Emotional Intelligence

employees' emotional labor tax but negatively related to their positive service behaviors. Furthermore, emotional labor tax partially mediates the relationship between problem customer perceptions and positive service behaviors. Results also reveal that high emotional intelligence buffers the positive inf1uence of problem customer perceptions on emotional labor tax but negative inf1uence on positive service behaviors. Finally, the implications, limitations and suggestion for future research are discussed.

Keywords: Conservation of resources theory; Emotional intelligence; Emotional labor tax; Positive service behaviors; Problem customer perceptions

1.

Introduction

Coping social stressor (Clayton, 1992; Dormann and Zapf, 2002; Brotheridge, and Lee, 2002) and problem customer (Hu, Hsu, Lee, Chang, and Hsu, 2011; Rupp and Spencer, 2006; Rupp, McCance, Spencer, and Sonntag,

2008; Spericer and Rupp, 2009) has been used by organizational researchers to refer emotional labor to regulate 企ontline employees' behaviors to display appropriate emotions on the jobs (e.g., Goodwin, Groth, Frenke, 2011; Gross,

2002; Hochschild

,

198; Scott and Bames

,

2011) to fulfill organizational requirements (e.g., Diefendorff, Croyle, and Gosserand, 2005; Grandey, 2003). More recently, these concepts has been employed to explain why frontline employees may result in emotional stress from emotional labor, so-called

“

emotional labor t的" when they struggle problem customers, who violate generally accepted norms ofbehavior during service encounters (H吼叫 al.叮 2011),

and then less on delivering willingly positive service behaviors (Harris and Reynolds, 2003; Judge~ Woolf, and Hurst, 2009).

In general, studies have emphasized the value of positive service behaviors-treating customers nice旬， having a kind smile, and providing the servlce ln 企iendly manners in service encounters (Trougakos, Beal, Green, and Weiss, 2008), such as. eliciting higher levels of customer satisfaction (Brown and Sulzer-Azaroff, 1994), customers' willingness to retum and recommend to others (Ts缸， 2001), and encouraging customers to spend more time in stores (Tsai and

Chiao Da Management Review Vol. 33 No.l, 2013 3

Huang, 2002). Though some studies on service encounters have focused on frontline employees' behaviors that promote customer satisfaction, however, for anyone who has ever coped with problem customers; it is sometimes hard to believe the old adage

“

the customer is always right"

,

despite the fact that service quality is a critical issue to organizations. The reality is that service encounters between 企ontline employees and problem customers may frequently result in negative and even depressing emotions (e.g., Bitner, Bernard, and Mo尬， 1994; Fullerton and Punj, 1993; Hu, et. al., 2011). Hu, et al. (2011) explained that the

higher frequency of interactions with problem customers may heighten 企ontline

employees' emotional labor burden and work-family conf1ict, because such interactions involve the need for 企ontline employees to fake/regulate their emotions in a mandated way.

Obviously, problem customer in service encounters is regarded as a source of social stressors,“a c1ass of characteristics, situations, episodes, or behaviors that are related to psychological or physical strain and that are somehow social in nature" (Dormann and Zapf, 2004: 62), that may cause negative consequences. As such

,

it is important to understand frontline employees' reactions to cope with problem customers, as well as the influences of coping problem customers lead to negative work outcomes 'to organizations and their subordinates and are worthy of continuing academic and practical concerns.

To expand knowledge in the area

,

we propose two goals in this current study. First, adopting conservation ofresources (COR) theory (Hobfoll, 1989, 2001) as a guiding 企amework， we focus to replicate past research linking problem customer perceptions to frontline employees' emotionallabor (Hu

,

et. al.

, 2011; Rupp and

Spencer, 2006; Rupp, et. al., 2008; Spencer and Rupp, 2009) and then examining the previously untested relationship between problem customer perceptions and positive service behaviors. We also like to investigate whether burden of

“

emotional s仕的s" resulted 企om emotionallabor

,

so-called

“

emotional labor tax"

,

potentially mediates the linkage between problem customer perceptions and positive service behaviors. The second goal of this study is to further examine the extent to which 企ontline employees' emotional intelligence can act as a buffer against the negative influence of coping problem customers.

of the Moderating ξffect of Emotional Intelligence

intelligence, defined as

“

an individual's capacity to process emotional information to enhance cognitive activities and facilitate social functioning" (Rivers, Brackett,

Salovey, and Mayer, 2007: 230), has been shown to be a positively reliable predictor of a number of important workplace outcomes (Zeidner, Matthew, and Roberts, 2009). As COR theory suggests, social stressors are thought to reduce employees' overall coping abilities as they drain important emotional resources (Hobfoll, 1989, 2001). However, COR theory suggests that 企ontline employees with higher levels of emotional coping abilities are better able to resist the negative effects of stressors. In this study, we posit that emotional intelligence can have a similar buffering effect, due to its favorable inf1uence on coping emotional resources. However, we also like to reason that COR theory can be used to argue the relevance of emotional intelligence as a moderator.

2. Theory and Hypotheses

2.1. Problem Customers

In exploring and describing customers to disrupt service in a manner, a variety of terms and phrases have been employed

,

including

“

deviant consumers" (Mills and Bonoma, 1979),“aberrant consumers" (Fullerton and Punj 1993), 吃jaycustomers" (Lovelock, 1994),“inappropriate customers" (Strutton, Vitell, and Pelton, 1994), “dysfunctional customers" (Harris and Reynolds, 2003),

“

interactional injustice customers" (Rupp and Spencer, 2006), and

“

problem customers" (Hu, et. α孔， 2011). In this study, we define the term “problem customers" 企om frontline employees' perspectives so that it refers to the degree of an employee 's perception of his or her customers who deliberately or unintentionally disturbed servicers that negatively affected the organization or other customers (Hu, et. al., 2011). Due to the high 企equency of interpersonal interactions between customers and frontline employees in service encounters,

thus, coping problem customers is often the most bothersome part of service jobs. Problem customer perceptions are inherently subjective assessments for frontline employees, and are primarily considered as a source of social stressors.

Chiao Da Management Review 均1.33No.1, 2013 5

2.2. Emotional Labor Tax

Organizations in service settings undoubtedly “commercialize" 企ont1ine

employees' emotions as a means to at仕a仗， please and retain customers (Hochschi1d

,

1983). The dramaturgical view of the service encounter is that

企ont1ine employees feel they must put on an act to present appropriate emotions to their customers (Goffm徊， 1959; Hochschild, 1983). This is a form of emotionallabor, wherein an employee manages their feelings to create a public1y observable facial and bodily display to perform organizational requirements and expectations (e.g., Diefendorff, Croyle, and Gosserand, 2005; Grandey, 2003), as well is performed t尬。ugh two identified forms (or s仕的gies): surface acting (involves the suppression of emotions and then fake emotional displays) and deep acting (attempts to acωally feel the emotions and then express a desired emotion) (Gross, 2002; Holdschild, 1983). Both surface acting and deep acting requires

企ont1ine employees to intentionally fake/modi砂 their own emotions. Studies indicated that frequent use of surface acting is generally related to negative job outcomes and deep acting can result in positive job outcomes (e.g., Brotheridge and Lee, 2002; Grandey 2003).

Usually，企ont1ine employees experienced negative emotions as coping problem customers; they may have lower motivation to work. Consequent1y, they might choose to adopt more surface acting because it only involved

“

superficial" changes of extemal emotions (e.g. hiding negative emotions). Furthermore,

individuals may also result negative job outcomes as often masking their true fee1ings and incurring as a result of psychological stress of emotional taxing 企om

deep acting

,

because these positive outcomes of deep acting may be offset by the negative emotional experience that originates from coping problem customers. That is to say that frontline employees heighten emotional labor tax on emotional acting under implicit and explicit pressure to fake/regulate their emotions as a tool to attract customers into patronizing the organization (e.g., Brotheridge and Grandey, 2002; Grandey, 2003), specifically as coping problem customers. In this study, emotionallabor tax, which differs from deep acting and surface acting (即o

01 the Moderating Effect 01 Emotional Intelligence

individual's burden of

“

emotional stress" resulted 企om deep acting or surface achng.

2.3. Conservation of Resources (COR) Theory

A review of the literature indicated that conservation of resource studies have been conducted by both Lazarus and Folkman's (1984) and Hobfoll's (1989). Basically, an assumption under1ying the conservation of resources is that individuals strive to protect and conserve the resources they value (Hobfoll

,

1989; Lazarus and Folkman, 1984). Hobfoll defined resources

“

as those objects,

personal characteristics, conditions, or energies that are valued by the individuals or that serve as a means for attainment of these objects" (Hobfoll, 1989: 516). And Lazarus listed as resources the following characteristics

,

including intelligence

,

money

,

social skills

,

education

,

supportive family and friends

,

physical attractiveness

,

health and energy

,

and sanguinity (Lazarus

,

2001 :382). While the substance of Lazarus' critique of conservation of resources (COR) theory is examined closely to emphasize that

“

Hobfoll' s viewpoint is fundamentally unsoùnd and fails to advance beyond what he know" (Lazarus,

2001: 381). For all that, there are still 制TO reasons to explain why Hobfoll's COR theory is much more relevant in this study than Lazarus's appraisal theory and then we adopt COR theory as a theoretical foundation in this sωdy.

Firstly, COR theory implies that individuals make efforts to conserve their valued resources, so as to better attain desired goals (Grandey and Cropanzano,

1999). According to COR that stress occurs when individuals are either threatened with resource loss, actually lose resources, or fail to gain resources following a significant resource investment (e.g., coping problem customers). This loss of resources can lead to a negative state ofbeing

,

namely

,

strain (Hobfoll

,

1989). But Lazarus (1984) addressed that stressor comes 企om stimuli

,

in which leads to psychological distress, or physical impairment or deterioration. In this current study, we argued when individuals under stressful conditions (i.e., coping problem customers) may experience emotional resources loss rather than the stimuli, they may lead to the burden of

“

emotional stress" resulted from deep

Chiao Da Management Review Vol. 33 No.1, 2013 7

acting or surface acting in order to perform expected emotions to customers (Hochschild, 1983).

Secondly

,

the concept ofresource loss is central to Hobfoll (1989) approach in the experience of stress and coping. Whereas

,

according to Lazarus (1991)

,

stress appraisal model is the evaluation of gains and losses with regard to the wel1-being of the person. In other words, COR theory suggests that perceiving

s仕essors (i.e., coping problem customers) and being able to handle the problem motivate action (e.g., emotional regulation). Accordingly, COR theory 企om

Hobfoll (1989) can examine a relationship between stressful person and environment (e.g., coping problem customers). But Lazarus (2001) addressed that primary appraisal is the evaluation of whether something relevant to our well-being has occu叮ed (Lazarus, 1991; Lazarus and Folkm妞， 1984). Therefore,

stress appraisal model 企om Lazarus widely applied in the depression literature (Lazarus, 2001), and COR theory 企om Hobfol1 suitably related to emotional regulation literature (Brotheridge

,

and Lee

,

2002).

Social stressors duritig service encounters, such as problem customers may threaten and/or cause a depletion of valued emotional resources. Thus, coping problem customers being viewed as a s甘essful work status in service 仕ansactions

(i.e., conditional emotional resources loss), experiencing feelings of service failure (i.e.

,

personal emotional resources loss)

,

or spending efforts coping with bad interpersonal situations (i.e., emotional energy expending) all can deplete emotional resources, resulting in emotional taxing 企om emotional labor (Hu, et.

α1. ， 2011) and then less on delivering willingly positive service behaviors (Judge

,

et. a孔， 2009).

2.4. Problem Customer Perceptions

,

Emotional Labor Tax and

Positive Service Behaviors

Studies have examined the impact of customer interactional injustice perceptions on emotionallabor (e.g., Rupp and Spencer, 2006; Rupp, et. al., 2008;

Spencer and Rupp, 2009), which found that customers may communicate and interact with 企ont1ine employees in an uncivil or a discourteous manner. Anger and unhappiness are cited as the most common negative emotions experienced by

Dance with Problem Customers: An Examination

0/

the Moderating ξffect

_0/

Emotional Intelligence

front1ine employees in such situations (Clayton

,

1992). The result is that frontline employees when coping problem customer find it more difficult to comply with emotional display rules than well-treated employees by customers. They thus tend to engage in heavier emotionallabor in order to adhere to organizational display rules (Ashforth and Humphrey, 1993; Diefendor匠， et. a孔， 2005).

As noted above, coping social stressors (i.e., problem customers) during service encounters are thought to promote a number of undesirable outcomes for individuals and organizations. In this study, we like to rely on the COR model (Hobfoll, 1989, 2001) to explore the relationship between problem customer perceptions and emotionallabor tax. That is to say that it would deplete personal emotional resources as well as threaten the loss of emotional energy in coping with problem customers for competing demands of work, therefo間，企ontline

employees generally have to put more efforts into faking/regulating their emotions in order to follow organizationally-desired expectations

,

and may therefore heighten emotional labor tax (emotional s仕ess) to mask their true feelings (Ashforth and Humphrey, 1993; Diefendor缸~ et. al., 2005). The labor of inhibiting felt or“true" emotions requires a great deal of physiological effort (e.g.,

Goodwin, Groth, and Frenke, 2011; Gross and John, 2003; Scott and Bames,

2011) of emotional suppression (e.g., Brotheridge and Lee, 2002), and invests a significant amount of emotional labor to maintain relationships with problem customers and other customers at work. The higher 企ont1ine employees perceive problem customers

,

therefore

,

the higher they load emotional taxing resulted 企om

emotionallabor. Thus, we propose to test the following hypothesis:

Hl: Problem customer perceptions are positively related to 戶。ntline

employees' emotionallabor tl肌

Specifically, based on Hobfoll (1989) COR theory, it is argued that people strive to protect and retain resources under stressful conditions. Psychological stress is sparked when there is a threat of net loss of resources, a net loss of resources actually occurs, or when resources are not gained following the investment of resources (Hobfoll, 1989). However, COR theory of stress argues that certain behaviors and attitudes are likely to occur as a result of emotional

ChiaoDα Management Review Vol. 33 No.l, 2013 9

resources loss. As COR theory suggests, social s仕essors are thought to reduce employees' overall emotional coping abilities as they drain important resources (Hobfoll, 1989, 2001). Therefo時， individuals under stressful conditions (e.g., coping problem customers), one potential behavioral reaction, in order to maintain a balance between the demands placed upon them and their emotional resources, is to attempt to reduce work demands (e.g., less positive service behaviors) (e.g., Judge, et. al., 2009) in an effort to protect one's emotional resources during service encounters. We therefore offer to test the following hypothesis:

H2: Problem customer perceptions are negatively related to frontline employees ' positive service behaviors.

In this study, we conceptualized emotional labor tax as a mediator between problem customer perceptions and positive service behaviors. According to COR theory, individuals lose emotional resources when they encounter emotional stress (Hobfoll

,

1989). In otber words

,

more negative perceptions of one's job situation will more likely lead to higher perceptions of resources being threatened (Hobfoll, 1989). Coping problem customers (i.e., social stressors) in service encounters are the negative perceptions of 企ontline employees' job situations. Such s甘essors

adversely a宜ect people by investing a significant amount of emotional labor to suppress emotions for maintaining relationships with customers at work, and then by attempting to reduce work demands (e.g., less positive service behaviors) in an effort to protect one's emotional resources in service encounters. Based on the above argument of Hypothesis 1 and 2, since service organizations always ask

企ontline employees to put on an act to present appropriate emotions. In this s切dy，

we argue that when 企ontline employees encounter problem customers

,

they may increase employees' emotional stress from emotionallabor because they are asked to create a public1y emotional display to satis勾r organizational requirements (e.g.,

Diefendor缸， et. α1. ， 2005; Goodwin, et. al., 2011; Grandey, 2003; Hochschild, 1983; Scott and Bames, 2011). More specifically, heighten emotional stress resulted from emotionallabor would eventually relate to weaken their willingness to deliver positive service behaviors. Hence, we like to test the following hypothesis:

01 the Moderating ξffect 01 Emotional Intelligence

H3: Frontline employees' emotional labor tax mediates the relationsh息，

between problem customer perceptions and positive service behaviors.

2.5. Emotional Intelligence (EI)

The term emotional intel1igence-which we abbreviate to EI-has been

ascribed to various sources and is a fair1y popular cons甘uct in organizational

research. In that short time it has created much discussion and controversy among

researchers and practitioners. However, the first systematic research on EI was

conducted by J ack Mayer and Peter Salovey. They define EI as a col1ection of

"abilities to monitor one's own and others' feelings and emotions, to discriminate

among them, and to use this information to guide one's thinking and actions"

(Salovey and Mayer, 1990). Later, they refined their definition as the ability to

perceive emotions, to access and generate emotions so as to assist thoughts, to

understand emotions and emotional knowledge, and to reflectively regulate

emotions so as to promote emotional and intellectual growth (Mayer and Salovey,

1997). To Goleman (1995) EI is "the capacity for recognizing our own feelings

and those of others, for motivating ourselves, and for managing emotions wel1 in

ourselves and in our relationship". He focuses EI on performance at work and

organizationalleadership, and depicts EI as a model of emotional competencies:

self-awareness, self-management, social awareness, and relationship management

(Zeidner, et. al., 2009). Moreover, for the measurement of 凹， Bar-On in仕oduced

the term emotional quotient (EQ) and defined EI as an array of emotional and social knowledge and abilities that influence our overall ability to effectively copy

with environmental demands (Bar國On， 2000).

Since EI has been defined and used by different researchers in various ways,

in this study, we adopt an integrated four-dimensional definition and a self-report

scale of EI proposed by Wong and Law (2002), which was developed based on

the definition of Mayer and Salovy (1997). Wong, Law and Wong (2004) argued

that this self-reported scale of EI measure is valid for Chinese and they proposed

fo凹 dimensions: self emotional appraisal (individual's ability to understand their

deep emotions and enable to express their emotions naturally), others' emotional

Chiao Da Management Review Vol. 33 No.1, 2013 11

people around them), regulation of emotion (the ability of individual to regulate their emotions, which can enable quick recovery from psychological distress), and use of emotion (the ability of individuals to make sure of their emotions by directing them towards constructive activities and personal performance) (Wong and Law

,

2002; Wong

,

et. al., 2004).

2.6. The Moderating Effect of Emotional Intelligence (EI)

Employing COR theory as the guiding my reasoning

,

we argue that higher EI can act as a buffer against the undesirable impact of problem customer perceptions on emotional labor tax and positive service behaviors

,

respectively. Specifically

,

to the extent that problem customer perceptions heighten heavier tax on emotional stress from emotional labor and then diminish positive service behaviors by depleting individuals' emotional resources，企ontline employees with high levels of coping emotional resources will have a greater tolerance for problem customers than those with low levels. As noted by Hobfoll (1989), high,

as opposed to low, levels of coping emotional resources should theoretically permit individuals to endure relatively high levels of social stressors (i.e., coping problem customers).

We argue that EI is relevant to this discussion due to the abilities to act as coping emotional resources. The notion of EI can function in this way is based on three lines of reasoning. First, stressful events are pressures or constraints that challenge one's emotional coping abilities. In fact, coping involves a person's efforts to manage the demands of a person-social transaction that is appraised as stressful (e.g., Folkman, 1991). These constantly changing efforts can be cognitive or behavioral

,

direct and indirect (Zeidner

,

et. al.

,

2009). Accordingly

,

when the demands of an acωally stressful situation (e.g.

,

coping problem customers) are perceived as stressful, efforts are directed at regulating emotional stress (i.e.

,

regulation of emotion) (Wong and Law

,

2002). Epstein (1998) indicated that higher EI individuals may have richer emotional and social coping resources compared to their lower EI counterparts. Thus, we expect that here is a direct link between EI and a person's ability to cope with social stressors (i.e.,

01 the Moderating Effect 01 Emotional Intelligence

EI are relatively unlikely to take socially s仕essful events personally, given their relatively positive self-perceptions. Meanwhile, research suggested that EI plays an important role in emotion-related performance (Bechtoldt, Rohrmann, De Pater,

and Beersma, 2011; Mayer and Salovey, 1990, 1997; Yoo, and Salovey,

2008/2009). As such, there may be less of an emotional cost or tax associated with social stressors for these 企ontline employees, promoting higher levels of positive service behaviors

,

a greater ability to avoid emotional resources loss associated with spending less emotional resource and taxing less emotional acting in service encounters (Bechtoldt, et. al., 2011).

Second, COR theory suggests that positive aspects of the seIf appraisal will likely serve as coping resources when exposed to stress (Hobfoll, 2001). Two components of 凹， self emotional appraisal and others' emotional appraisal, are

likely to be served as the abilities. Specifically，企ontline employees with higher self's and others' emotional appraisal will have more resources available for coping with negative social interactions (Bechtoldt, et. al., 2011). These components are likely to provide employees with a sense of con仕01 to express organizationally-desired emotions naturally (high self emotional appraisal) or better enable them to react more calmly to perceive and understand the emotions of problem customers (high others' emotional appraisal), both of which will lead to less emotional resources being depleted from these negative workplace interactions (Hobfoll, 1989). Moreover, Mayer and Salovey (1997) argued that high EI may result in employees increasing their affective commitment to the organization by generating enthusiasm for their work, especially for their positive service behaviors. Therefore, we argue when coping problem customers during service interactions, frontline employees with high EI may find it easier to handle stress and regulate emotion (i.e., regulation of emotion), and motivate themselves,

and in turn view coping problem customers as an opportunity to express their identity to service awareness, and then perform in more positive service behaviors during service encounters (i.e., use of emotion) (Wong and Law, 2002; Wong, et.

al., 2004).

Third, COR theory suggests that a more positive perception of one's job or job situation will likely lead to lower percepti

Chiαo Da Management Review Vol. 33 No.1, 2013 13

(Hob fo 11 , 1989). Thus, we expect that the impact of higher EI at work might also promote emotional coping abilities that can act as a buffer against coping social stressors (i.e.

,

'problem customers) in service encounters. Wong and Law (2002) explained that high EI help individuals perceive their job in a positive light,

causing them to focus on the desirable characteristics of the jobs. In other words,

frontline employees with high EI enable quick recovery 企om psychological distress (i.e., regulation of emotion) and make sure of their emotions by directing them towards positive activities and personal performance (i.e., use of emotion). Therefore, we ar，邵le that high EI can serve 企ontline employees to increase their positivity regarding their jobs. The abilities can help 企ontline employees see negative social interactions (i.e.

,

coping problem customers) at work as isolated episodes. Additionally, Hobfoll (2001) noted that such positivity can act as an emotional resource for coping with social stressors. Therefo肉， we argue that higher EI can add to 企ontline employees' stocks of coping emotional resources by adopting less emotional acting to diminish emotionally taxing for frontline employees in service encounters.

Taken together, these arguments suggest that frontline employees with higher EI will have greater abilities to cope with problem customers than those with lower EI. Therefore, we argue that the heightened emotional coping abilities can attenuate the impact of problem customer perceptions on emotional labor tax by providing additional coping emotional resources (i.e., regulation of emotion) (Wong and Law, 2002) that can offset the depleted by coping problem customers,

and by performing less emotional acting in service encounters. The emotional coping abilities associated with higher EI are also predicted to weaken the negatively influence of problem customer perceptions on positive service behaviors by promoting high resource levels that can withstand the depleting effect of positive service behaviors. Frontline employees with high EI are also expected to see incidents of coping problem customers of social stressor as isolated events

,

and then are willing to deliver positive service behaviors in service encounters (i.e., use of emotion). Base on the above, we propose the following moderation hypotheses:

ofthe A必derating 珈ctof Emotional Intelligence

H4: Frontline employees' emotional intelligence 個刃 moderates the

positive relationship between problem customer perceptions and emotional labor t，缸.， such that the relationship is weaker for those with higher EI and stronger for those with lower EL

H5: Frontline employees' emotional intelligence (EI) moderates the negative relationship between problem customer perceptions and positive service behaviors

,

such that the relationship is weaker for

those with higher EI and stronger for those with lower EL

3. Method

3.1. Participants and Procedures

Since the settings of this study are service contexts. In order to make an appropriate sample for this study of the concepts of problem customer perceptions and emotional labor tax during service encounters, participants were recruited

企om full-time 企ontline employees of restaurants. Due to the high frequency of interactions between customers and 企ontline employees in restaurants, restaurants are deemed appropriate for this study because there is general a high proportion of

employed 企ontline employees may cope with problem customers and all the time

conforming to the organization's emotional display rules. In this study, we conducted over a period of approximately half year in three separate located in the cities of Taip缸， Taichung and Kaohsiung in Taiwan and distributed 600 matched-pair questionnaires to 80 mid-size restaurants (at least 100 capacities or seats for customers in restaurants).

Since previous research suggested that a same-source bias may inflate actual relationships (Podsakoff, MacKenzie, and Podsakoff, 2003; Podsakoff and Organ,

1986), using different sources helps alleviate the potential same-source bias. Therefore, we adopted an in-dyad research method in this study. In other words,

the questionnaires were administered to both frontline employees and their immediate supervisors

,

and were completed during work hours. The surveys were

Chiao Da Management Review Vol. 33 No.1, 2013 15

performed at two different time periods with a two-week time lag in order to

minimize the potential common-method bias too. At Time 1, participants (i.e.,

frontline employees and their immediate supervisors) received questionnaires and

postage-paid envelopes at their work addresses

,

together with a covering letter

explaining the purpose of study and inviting them to participate. The letter also

emphasized the importance of honesty in responding, and guaranteed anonymity

and confidentiality of their answers. Participants were asked to provide their name

so that we could match data collected at later times. However

,

respondents were

assured their names would be used for research purposes only and that all identifying information would be removed after data were coded. After

completing the survey, participants were asked to return the completed

questionnaires in the pre-paid envelope. After the data were collected 企om

企ontline employees at Time 1

,

supervisors were then asked to rate their respective

employees at Time 2

,

two weeks after Time 1 data were collected.

To ensure confidentiality and to diminish socially desirable responses

,

企ontline employees and their immediate supervisors were informed that their participation was voluntary and that a coding system was used to match the

surveys. Accordingly, we coded each questionnaire with an assigned

identification number to match 企ontline employees' responses with their

immediate supervisors' evaluation. The questionnaire for frontline employees

comprised four parts: demographics

,

problem customer perceptions

,

emotional

intelligence, and emotional labor tax. Additional妙， emotional labor is a

requirement and expectation of organizations, and supervisors are responsible to

ensure that 企ontline employees fulfill this requirement. Therefore

,

we argue that

supervisors should be the most appropriate people to judge emotional stress

resu1ted 企om emotional labor of 企ontline employees. In order to minimize the

potential bias in results, immediate supervisors completed the assessment of

positive service behaviors and emotional labor tax for their matched frontline employees.

We discarded 11 returned questionnaires with excessive missing data and 6

without accurately matched as dyads. After deleting invalid responses, we

with Problem Customers: An Examination 01 the Moderating Effect 01 Emotional Intelligence

of 39%. Among the participants of 企ontline employees, 58% were female, over half (61 %) had at least a college degree

,

53% were at least 25 years old

,

and 630/0 had worked in restaurants for over a year. Of the supervisors participants, 87% were male; the average age was 34 years (S.D.=4.83); the average organizational tenure was 3.1 years (S.D.=2.50); and they were generally well-educated, with 840/0 having completed a college or university degree. Since we explore 企ontline employees' perceptions and their behaviors reactions, therefore, only the demographics provided from employees were used to data analyses in this study.

3.2. Measures

This study followed Brislin's (1980) translation-back-translation procedure to create a Chinese version of the questionnaire containing measures of focal variables, except for the problem customer perceptions scale. The measures used in this study were translated 企om English into Chinese by a professional

甘anslator. The Chinese versions ofthese materials were then back-translated 企om

Chinese to English by a Chinese doctoral candidate f1uent in both English and Chinese. The student did not see the original English version. Both of the original materials and their reverse-translated counterparts (English version) were evaluated by a bilingual psychologist to ensure that the translations were correct and that the content was the same.

Problem Customer Perceptions (PCP). To measure problem customer perceptions, based on the existing scales 企om Colquitt's (2001) along with the discussions with industry and academic professionals

,

eight items were developed specifically for this study as shown in Appendix. We employed

“

my customer" 的

the referent in each item. To avoid socially desirable response，企ontline

employees were asked to indicate the degree to which customers had treated them

a~ violating generally accepted norms of behaviors during service encounters, . as described on a 5-point Likert scale ranging from 1 (not at all) to 5 (very great

extent). This meant that higher scores denote a higher degree ofproblem customer

perceptions. The exploratory factor analysis revealed a one-factor structure, and alpha re1iability was measured as .83.

ChiaoDaManα'gement Review Vol. 33 No.1, 2013 17

Emotional Labor Tax (EL T). The emotional labor tax scale, consisted of a six-item self-report measure adapted from Brotheridge and Lee (2003). The six-item questionnaire reflected the extent to which the individual's burden of emotional s仕ess 企om deep/surface acting offaking/regulating emotions as service part of the work role. Frontline employees were asked to indicate the extent to which emotionallabor tax they engaged in display appropriate emotional acting to perform their service jobs effectively during service encounters. The items are:

“

'During service encounters, 1 hide my true feelings about a situation",

“

During service encounters, 1 resist expressing my true feelings", and

“

During service encounters, 1 pretend to hαve emotions that 1 don 't really haνe." (surface acting) and

“

'During service encounters, 1 make an 壁沙ortto actually feel the emotions that 1 needed to display to others"

,

“

'During service encounters, 1 tη1 to actually experience the emotions that 1 need to display to others", and

“

'During service encounters, 1 real抄的Fω feel the emotions 1 have 的 show as part of my job."

(deep acting). The answer format using a 5-point scales: 1 =very lightly or not at all

,

2=a little

,

3=moderately; 4=quite a bit; 5=extremely. However

,

in this study

,

we combined the two components into a composite measure of emotional labor tax to have a conceptual justification for examining a total level of tax resulted

企om emotional labor as the dependent measure than each of its separate components. Exploratory factor analysis for ELT revealed a one-factor structure. Alpha reliability was measured as 0.85. Due to we cannot conclude all the items to be

‘

observable' only by 企ontline employees, but we argue that supervisors should be the most appropriate people to 'judge' emotional display of 企ontline

employees. Therefore, 6-item to be judged frontline employees' emotionallabor tax by supervisors were the same as those for the self-reported scale of employee rating (Brotheridge and Lee, 2003) we mentioned above, except that the former were referenced to

“

my subordinαte" rather than

“

f\All items were scored on a 5-point scale (1 =very lightly or not at all, 2=a little, 3=moderately; 4=quite a bit; 5 =extremely) , and the coefficient alpha of these six items was .80. Finally, we calculated to average on scales consisting of employee ratings and supervisor judgments

,

which we labeled emotional labor tax for

01 the Moderating Effect 01 Emotional Intelligence

Emotional Intelligence (EI). A total of 16 items, adopted from Wong and Law (2002) and Wong, et al. (2004), were used to measure 凹， consisting of self-emotion appraisal, others' emotional appraisal, regulation of the emotion, and use of emotion. Each of the four subscales consisted of four items. The 1師6ι創-寸品i扯temn 1

reflected s倪叫elf臼L冶emo仰t位10∞n appraisal ('‘ '1 have good understanding 01 my own

emη'lotio仰n.'六")，

emotions 01 01的her，付's.'六"勻')， regulation of the emotion (“1 am able to control my temper

and handle 荷伊culties rationαlly.")， as well as use of emotion ("1 always tell

myself 1 am a competency person."). All items are presented in Appendix.

Front1ine employees reported their own EI, each rated on a 5-point Likert scale ranging from 1 (not true ofme) to 5 (true ofme). The exploratory factor analysis yielded a four-factor structure, and the respective alpha reliabilities were .79, .74, .81, and .80. In this study, we combined the four dimensions into a composite measure of EI in order to have comprehensive conceptual justification for examining EI as the moderator measure than each of its separate dimension. Thus, single factor subscales consisting of all 16 items, which we labeled emotional intelligence for analytic pu中oses， produced an overall intemal reliability coefficient of .78.

Positive Service Behaviors (PSB). Immediate supervisors were asked to rate the selected employees' positive service behaviors during service interactions. A three-item scale, adopted from Tsai and Huang (2002), was used to measure positive service behaviors, assessing employees' interactions with customers based on a 5-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly

agree). The items are:

“

My subordinate has a kind smile during service

interaction" ,

“

My subordinate provides the service in a friendly manner.", and

“

My subordinate treats customers nicely." The exploratory factor analysis

revealed a one-factor structure

,

and alpha reliability was measured as .84.

Control Variables. We controlled the front1ine employees' gender, age,

education, and job tenure because of their potential relevance to the independent and dependent variables, based on past research (e.g., Hochschild, 1983; Wharton and Erickson, 1993). For example, women are expected to do more eniotion management than men (Hochschild, 1983), as well as emotional1abor (Wharton

Chiao Da Management Reνiew Vol. 33 No.l, 2013 19

and Erickson, 1993). In addition, age, education, and job tenure were considered important con仕01 variables, because older, higher educated, or more experienced

企ontline workers are more adept at controlling their distress and emotions

,

and create necessary feelings and display appropriate emotions to promote a stronger job performance (Hochschild

,

1983).

Actually, discriminate validity verifies that measures of di任erent constructs are unique. Thus, CF A is allowed to examine whether the participants conceptualized of the focal constructs in this study as distinct and separate. We examined two models: a one-factor model comprised of all items in all four scales and a four-factor model forcing each item to load on its intended factor. The fit statistics for the one-factor model were root mean square error of approximation (RMSEA) of .20, comparative fit index (CFI) of .51, goodness of fit index (GFI)

。f .72

,

and Chi -square (χ，2) of 896.71 (p < .001). For the four-factor model

,

RMSEA = .052, CFI = .94, GFI = .鈣， and X2 = 556.68 (p < .001). Acco必ng to

Hu and Bender (1999) the RMSEA should be lower than .06; the CFI should be above .95 for good fit. With these cutoffs in mind, we considered that the fit of the four-factor model was acceptable. Besides, the Chi-square difference test resulted in a change-in-Chi-square (I1X2) of 175.12 (p < .001). This suggested th的 the four- factor model fits the data significantly better than the one-factor mode1.

The statistically significant I1X2

,

the nearly acceptable fit statistics of the four-factor model

,

and unacceptable fit statistics of the one-factor model suggested that our respondents conceived of the four cons甘ucts as distinct and separate.

Furthermore, two-tailed tests of significance for the relationship between the demographic variabl~s and the focal constructs were run. The correlations between the latent variables were all within a low to moderate range indicating good discriminate validity (Kline, 1988). Table 1 shows the means, standard deviations, and correlations of the study variables. Regarding the relationships among the focal cons甘ucts， all were significantly related at p < .01, the positive correlation between ELT (employee rating vs., supervisor judgments, respectively) and PCP (戶 .23 ， .20, p < .01, respectively) suggests frontline employees with high problem customer perceptions heighten emotional labor taxes in service

20

01 the Moderating ξffect 01 Emotional Intelligence

encounters

,

while the positive correlation between EI and PSB (r = 況 ， p < .01) suggests employees with high emotional intelligence display more positive service behaviors toward customers.

4. Results

In order to determine if significant relationships existed among the independent variable (i.e.

,

PCP) and dependent variable (i.e.

,

PSB) and the mediator (i.e., ELT), and were consistent with the method used to test mediating variables suggested by Baron and Kenny (1986). Due to the demographic characteristics would be related to problem customer perceptions, emotionallabor tax and positive service behaviors, we controlled for gender, age, education level and job tenure in each of regression models. The first regression analysis (Ml on Table 2) demonstrated that PCP 伊= .29; p < .01) explained significant variances in EL T, thus, Hypothesis 1 was supported.

The second regression analysis (M3 on Table 2) demonstrated that PCP explained significant variances in PSB (戶=眉之6; p < .01), Hypothesis 2 was also supported. A further regression analysis (M4 on Table 2) demonstrated that ELT explained significant variances in PSB 伊= .21; P < .01). Finally, to determine if a significant mediating effect existed

,

the changes in the size of the direct effect of PCP on PBS (M3 vs. M5 on Table 2) were observed. A partially mediating role was found to exist

,

because the size of the direct effect was reduced from -.26 (p

< .01, M3) to -.12 (p < .的， M5), when EL T was inc1uded (戶= .18; p < .05). Hypothesis 3 gained partially supported.

To test the proposed moderation hypotheses, we conducted hierarchical regression analyses. As noted above, certain demographic characteristics would relate to emotional regulation, we con仕ol1ed for gender, age, education, and job tenure in each of regression models. Table 2 states the statistical results for the moderation tests (M2 and M6). Fol1owing Aiken and West's (1991) suggestion,

the dependent variable on the control variables, independent variables, and the intera,ction term that are the components of the interaction term in the moderation analysis.

ChiaoDα Management Review Vol. 33 No.1, 2013 21

We regressed the dependent variable of EL T on the control variables (gender,

age, education level, and job tenure), independent variables (PCP), and the

interaction tenn (PCP x EI) (M2 on Table 2). The regression resulted in a

significant

F-scor

,

e

with an R-squared value of .16. The beta weight for PCP was

significant at .22 (p < .05), the beta weight for EI was also significant -.11 (p < .05)

and the beta weight for the interaction of PCP and EI was significant at 弋20 (p

< .01). In order to examine the nature and fonn ofthe interactions, we plotted the

interaction by developing separate equation using one standard deviation above and below the mean to represent high vs. low on each respective variable (Aiken

and West, 1991). Figure 1 depicts the interaction graphically for the interaction of

PCP and EI. The findings indicated that

,

indeed

,

the higher level of ELT happened

when f泊的line employee has lower EI and s甘onger PCPs. Conversely

,

the lower

level of ELT happened when 企ont1ine employee has higher EI and s甘onger PCPs.

Thus, Hypothesis 4 was supported.

In a similar vein, we regressed the dependent variable of PSB on the control

variables (gender, age, education level, and job tenure), independent variables

(PCP)

,

and the interaction tenn (PCP x EI) (M6 on Table 2). The regression

resulted in a significant

F-score

with an R-squared value of .31. The beta weight

for PCP was significant at -.24 (p < .01), the beta weight for EI was also

significant .29 (p < .01) and the beta weight for the interaction ofPCP and EI was

significant at -.32 (p < .01). We plo位ed the two-way interaction according to the

guidelines of Aiken and West (1991). Figure 2 shows that the pa仕em of the

two-way interaction is consistent with Hypothesis 5

,

indicating the higher PSB

happened when 企ont1ine employee has higher EI and s甘onger PCPs. Conversely,

the lower PSB happened when 企ont1ine employee has lower EI and s甘onger

PCPs. Hypothesis 5 therefore was supported.

5. Discussion and Implications

This study makes twO contributions to literature. First, despite the

theoretical and empirical studies have explored the differences in 企ont1ine

Dance with Problem Customers: An Examination

01 the Moderating ξffect 01 Emotional Intelligence

2006; Rupp, McCance, Spencer, and Sonntag, 2008; Spencer and Rupp, 2009),

problem customer perceptions (PCP) has rarely been examined in conjunction with emotional labor tax and positive service behavior (PSB). This study addressed the gap by examining the role of emotional labor tax as a mediator

PCP and PSB. Meanwhile,

22

study

tolerance can help front1ine employees cope with problem customers (Hu, et. al.,

knowledge of some individual-level

characteristics may still remain relatively underdeveloped. Thus, in this study, we

also examine the extent to which higher emotional intelligence (EI) buffer the

negative influence of problem customer perceptions (i.e.

,

social stressors). To our

knowledge, no prior study has looked closely at these potential mediating and

moderating effects to explain the processes by which 企ontline employees'

problem customer perceptions relates to their emotional labor tax and positive

service behaviors

,

as well as the extent to which front1ine employees' emotional

intelligence can act as a buffer against the negative influence of problem customer

percept1ons. dis甘ess indicated that empirical between stressor -buffering our however, 2011); Figure 2 4.A ρtv va u σ8 ••• 且 EI

Interaction Plot for EI as a Interaction Plot for EI as a

Moderator between PCP and PSB Moderator between PCP and ELT

ES 區宮值 S zaω 占百。 h ...-..一 ..一..._...-﹒…一-._--. MBHHOSZ 有 g 濁。 g 凶 Low High Problem Customer P 前 ceptions

Low Higb

Problem CustOl世 rPerceptio固

.---Hi唱igl刮t岫副訓叫h恤Int叫t叫叫eee叫ell恥吋l

i →一 1山A訓w En岫田叫叫:DtiO。叩叩叫Dn8

ChiaoDaÀ 必 nagement Re ν iew Vol. 33 No.1 , 2013 Table 1 Means

,

Standard Deviations

,

Correlations

,

and Reliabilities Variables Mean S.D. C 2 1. Gender (FEt 1. 38 d .92 2. Age (FE) 25 .1 1 f 4 .3 7 -.12* 3. Education Level (FE) 2.75 e .95 .02 .02 4. Job Tenure (FE) 1.29 f .74 .06 .08 5. Problem Customer Perceptions (Time 1) 2.86 2.06 -.03 .04 6. Emotional Labor Tax (FE) (Time 1) 3.01 .91 .16* .02 7. Emotional Labor Tax (I S)b (Time 2) 2.82 .75 .19* -.01 8. Emotional Labor Tax (FE+IS) (Time 1+2) 2.92 .96 .1 7* -.01 9. Emotional In telligence (Time 1) 2.69 1. 90 .22** -.03 10. Positive Service Behaviors (Time 2) 3.21 1. 38 -.13* -.07 a: Fron t1 ine employees (FE) reported their own emotionallabor tax. b: Imm ediate supervisors (IS) judged fron t1 ine employees' emotionallabor tax. C: S.D. = Standard Deviation d: The coding of gender: 1 =m ale; 2=female. 3 _{-.06 .05 .06 .08 .07 .08 -.05} e: The coding of education level: 1 =high school; 2=college; 3=bachelor; 4 =m aster; 5=Ph.D. f: Age and job tenure were provided in number of years. 4 _{.14* .07 .04 .05 .11* .08} g: Intemal consistency values (Cronbach's Alpha) were shown on across the diagonal parentheses. 〉 <05; 刁< .01 5 6 (83)g .2 3 抖 (85) .20** .2 7** .21** .2 3** -.04 -.09 -.2 8** .22** 23 7 8 9 10 (80) .25** (83) -.08 -.08 (78) .24** .26** .3 1" 正 84)

01 the Moderating Effect 01 Emotional Intelligence

Table 2

Results of Hierarchical Regression Analysis Emotional

Positive Service Behaviors

Variables LaborTax Ml M2 M3 M4 M5 M6 Controls Gender .14* .13* -.13* -.12* -.12* -.ll * Age -.06 -.04 -.01 -.00 -.00 .01 Education Level .04 .04 .03 .02 .01 .02 Job Tenure -.07 -.02 -.06 -.04 -.03 -.05

Problem Customer Perceptions .29** .22** -.26** -.12* -.24抖

Moderator: Emotional Inte1ligence -.11 * .29**

PCP x EI -.20** -.32**

Mediator: Emotional Labor Tax

.21 ** .18* (ELT=FE+IS) F 9.48*** 7.36*** 19.43*艸 16.11*** 14.05材* 17.48"* R2 .17 .16 .33 .30 .26 .31 A吐justedR2 .15 .14 .31 .28 .24 .29 * _{p < .05; P} ** _{< .01;} *** _{' p < .001}

Second, while consideration attention has been paid in the past to problem

customers research issues related to justice theory (e.g., Rupp, et. 訓， 2008) and

affective events theory (e.g., Rupp and Spencer, 2006), little empirical evidence

has been found to establish a direct relationship between problem customers perceptions and conservation of resources (COR) theory. Using COR as the

theoretical basis, we found support for both of the main effects hypothesized in

the 企amework. Specifically, we found a positive relationship between problem

customer perceptions and emotionallabor tax, and a negative association between

problem customer perceptions and positive service behaviors. From a COR

perspective, we suggest that a likely reason for these findings is that individuals

reporting higher levels of social stressors perceive (i.e., coping problem customers)

Chiao Da Management Review Vol. 33 No.l, 2013 25

expend additional resources (e.g., emotional energy) in service encounters. For example, when problem customers are present during service interactions,

企ont1ine employees spend effort thinking about how to rnanage and possibly avoid these situations. To do so, these front1ine ernployees rnust expend valuable emotional resources, thus resulting in greater s仕ains for performing emotional acting and less energy in delivering positive service behaviors. As a result, coping problem customers lead to higher levels of emotionallabor tax, and lower level of positive service behaviors.

Given that problern customers may sometime present during service encounters, and given that it is difficult to eliminate them, we examined the potential for EI to minimizelbuffer the negative impact of these stressors. We found evidence for the interactive effect of problem customer perceptions and EI on both emotional labor tax and positive service behaviors. These findings are important because it supports the idea th剖 although social stressors (i.e., coping problem customers) have an undesirable influence on job outcomes, however,

problem customers do not impact all 企ontline employees in the same way. In other words, relative to those with lower EI, frontline employees with higher EI appear better equipped with the cognitive resources needed to cope with problem custorners effectively. To explain these findings, COR theory suggests that front1ine employees have different coping emotional resources that influence their abilities to cope with stressful situations (Hobfoll, 1989, 2001). As previous studies have shown that employees with higher EI view the workplace through more optimistic lenses (e.g., Bechtoldt, et. al., 2011; Carmeli, Yitzhak, and Weisberg, 2009; Zeidner, et. al., 2009), it is likely that EI provides frontline employees with beneficial coping emotional resources. That 峙， those with relatively favorable self ernotional appraisal that are more positive appear to have additional buffers against emotional resource loss than those with self emotional appraisal that are more negative. The finding that higher EI can buffer the negative effects of problem customer perceptions elicits to mitigate the impact of workplace-interaction stressors. Thus, we discuss this topic in a further detailed manner in the directions for practical implications and future research.

26

5.1. Practical Implications

01 the Moderating Effect 01 Emotional Intelligence

We submit the findings have four practical implications. First, the statistics

results reinforce the notion that problem customer perceptions can be especially

potent drivers of emotional labor tax and positive service behaviors, explaining

23% and 15% of the variance in the respective outcomes, respectively. COR

suggests that the depletion of coping emotional resources (caused by coping

problem customers) is a likely source of this explained variance. As such, it

appears critical that supervisors work to minimize potential social s仕essors (i.e.,

coping problem customers) by establishing appropriate service norms against

aggresslon 企om problem customers. Although it may be important to minimize

the number of task-specific stressors frontline employees may face, it is generally

accepted that these stressors (problem customers) are almost never completely

avoidable.

Furthermore, emotion regulation can be defined as the act of controlling

one's inner feeling states and an external reflection of the expected state in order to produce a desired emotion display that is considered appropriate in a given

situation (Hochschild 1979). Consequently, organizations are increasingly willing

to direct and control how 企ontline employees presented themselves to others. In

other words, the images 企ontline employees created for customers and the quality

of interactions between employees. and customers have come increasingly under the con甘01 of management. However, problem customers always occurred in

service encounters

,

the interaction of problem customers and service staffs also

influenced emotional reaction,“was the customer always right?", it has become

commonplace in past but worthy to discuss. These results unveiled that organizations implementing management mechanism should concentrate on

encourage, training and rewarding their 企ontline employees in how to be effective

emotion display and customer service management. Although organizations

cannot control the behavior of their customers, managers can take steps to

mitigate customer interactional justice effects on staffs' reactions by extending more fair treatment toward them to compensate for instances of unfairness

ChiaoDaManα'gement Review Vol. 33 No.1, 2013 27

hold 甘ue for justice perceptions coming from supervisors as compensating for

perceptions of injustice emanating 企om the customers.

A third implication concems employees' training and selection. Employees'

emotional abilities, such as those inc1uded in the EI constn叭， are generally

assumed to be leamable (e.g., Bechtoldt, et. al., 2011; Mayer and Salovey, 1997; Zeidner, et. a

L,

2009). EI is often a better predictor of a person's success in the

workplace (Ashkanasy and Daus, 2002). For service indl指出的， therefore, the

development of abilities of emotional intelligence is seen as a key determinant of

positive service behaviors tòward customers, and therefore helps underpin

c。中orate success (Carmeli, et. al., 2009; Voola, Carlson, and West, 2004). As

such, it may be practical for supervisors to 甘Y to enhance and improve 企ontline

employees' EI. We have argued, however, that higher EI will help 企ontline

employees perceive their job in a positive light (Carmeli, et. al., 2009), causing them to focus on the desirable characteristics ofthe jobs when working in socially stressful environments. This study provided preliminary evidence suggesting the development of abilities of EI is a potentially effective way to manage emotions

and to foster employees' positive service behaviors. Therefore

,

service sector

companies should also increasingly consider EI abilities as a potentially important selection criterion. Since in situations where problem customers tend to exist at high levels, such as

“

people-work" jobs (e.g., Hu, et. α1. ， 2011) that involve a

large amount of interaction with 企ontline employees

,

we therefore suggest that

supervisors seek to hire individuals who possess higher EI. Such individuals are

likely to have the resources necessary to cope with the social stressors inherent in these types of jobs.

Last, interestingly, the study also found that the gender control variable

exhibited significant effects on emotionallabor tax and positive service behaviors,

respectively, supporting the suggestion that female 企ontline employees are more

adept than males at displaying appropriate emotions (e.g., Hochschild, 1983),

carrying a lower emotional labor tax and conveying higher levels of positive

service behaviors to customers (e.g., Deutsch, 1990; Hochschild, 1989; James,

1989; Rafaeli, 1989). For exainple, Deutsch (1990) and James (1989) indicated

28 Dance with Problem Customers: An Examination 01 the Moderating Effect 01 Emotional Intelligence

the findings of the study also have implications for hiring practices in the service sector.

5.2. Limitations and Future Research

This study has certain limitations that also need to be taken into considerations. The sample population only consisted of frontline employees of restaurants in Taiwan and the overall sample size was relatively small (232 frontline employees). Meanwhile, various restaurants may operate under same or different management models. Therefore, the results may be somewhat limited in their generalizability. This issue could be addressed in future research with a larger sample of 企ontline employees across a broader range of service sectors.

We explore in this study about employees' perceptions of their struggling on problem customers

,

who unintentionally or deliberately disturbed servicers during service encounters (H吼叫仗， 2011). In fact, problem employees may also cause problem customers during service encounters. Despite it is not within the scope of this study to provide an extended discussion of the ongoing debates. Future study should be alerted this limitation and hope to clarify the concem for the judgments of

“

problem customers".

Evidence have examined a causal relationship between personality and emotional intelligence (e.g., Landy, 2005), and reported the influence of negative affectivity on emotional labor (Thoresen, Kaplan, Bars旬， Warren, and

deChermo帥， 2003). Future researchers may inc1uded these controls (i.e., personality and negative affectivity) based on prior research in order to gain a better understanding in relation to problem customer perceptions, emotionallabor tax, emotional intelligence, and positive service behaviors.

Similar to other studies on emotion management, this study used a cross sectional design to test the relationships among the variables. Even so, the design was still limited in being able to assert a causal relationship among problem customer perceptions, emotional labor tax, emotional intelligence, and positive service behaviors. To clarify the directionality of these relationships, it is recommended that 如何re research incorporate a longitudinal design to further confirm the causality of the aforementioned relationships.

Chiao Da Management Review Vol. 33 No.1, 2013 29

In this study, though the overall results are consistent for these outcome variables, however, we suggest that future research investigate the potential for higher EI to influence relationships between problem customer perceptions and other outcomes

,

such as job performance and motivation. Such research could add additional insight into the abilities of higher EI to increase overalllevels of coping emotional resources. Moreover, a suggestion for fu仙re research is to examine other variables that may buffer or minimize the negative impact of problem customer perceptions. In conc1usion, we believe that there is much to leam about the implications of problem customer perceptions and the abilities of EI to act as an interactive variable. We hope that the findings can serve as a foundation for

臼ture research and we believe that the use of COR provides the appropriate reasoning to clarify these relationships.

Appendix

Problem Customer Perceptions

1 feel my customer treats me in impolite manners. 1 feel my customer 仕eats me with insult.

1 feel my customer persists 企om improper remarks or comments. 1 feel my customer 仕eats me without dignity.

1 feel my customer c1arifies hislher service needs incompletely. 1 feel my customer explains hislher service needs unreasonable.

1 feel my customer has been unfair in hislher communications with me. 1 feel my customer informs hislher service needs in an untimely manner.

Emotional Intelligence

Se

lf-

emotion appraisal

1 have a good sense of why 1 have certain feelings most of the time. 1 have good understanding of my own emotions.