10
Emotional dissonance and emotional exhaustion among hotel employees in Nigeria Osman M. Karatepe *, Kayode Dare Aleshinloye 1 School of Tourism and Hospitality Management, Eastern Mediterranean University, Gazimagusa, Turkish Republic of Northern Cyprus, via Mersin 10, Turkey 1. Introduction In a turbulent business environment of scarce resources, intense competitive pressures, and rapid rates of technological change, hospitality firms aim to focus on flawless service quality to gain and sustain competitive advantages for success and survival. According to the resource-based view, inimitable human resources (e.g., highly qualified and meritorious frontline or customer– contact employees) serve this purpose. Frontline employees are in frequent face-to-face or voice-to- voice interactions with customers and do emotional labor, which is defined as ‘‘the effort, planning, and control needed to express organizationally desired emotion during interpersonal transac- tions’’ (Morris and Feldman, 1996, p. 987). There are three modes of acting: surface, active deep, and passive deep (Hochschild, 1983). Surface acting occurs when employees fake their emotions by changing their outer demeanor to conform with the organizational display rules while the inner feelings remain unchanged (Zapf et al., 1999). Active deep acting occurs when employees have to spend effort to manage emotions by regulating their expressive behaviors and inner feelings, and passive deep acting occurs when employees may automatically feel emotions needed for a particular situation (Zapf et al., 1999). Emotional dissonance, which refers to incongruence between felt and displayed emotions (Heuven and Bakker, 2003), captures Hochschild’s (1983) two modes of surface acting and passive deep acting, which are the opposite ends of a continuum (Kruml and Geddes, 2000). Employees in frontline service jobs are susceptible to emotional dissonance (Phillips et al., 2007). Such employees are also confronted with emotional exhaustion, which is a form of work- related strain and refers to the lack of energy and depletion of emotional resources due to excessive psychological demands (Boles et al., 2000). Emotional dissonance and/or emotional exhaustion may give rise to negative job outcomes such as poor job performance, job dissatisfaction, reduced organizational commitment, absenteeism, and turnover (Abraham, 1998; Cro- panzano et al., 2003; Grandey et al., 2004; Karatepe and Uludag, 2007; Lewig and Dollard, 2003; Morris and Feldman, 1997). Besides, there are ample tangible and intangible costs associated with occupational stress and emotional exhaustion/burnout in a world of increasing global competition (Gabriel and Liimatainen, 2000; Kelloway et al., 2008; Spielberger et al., 2003). Against this backdrop, this study develops and tests a model, which investigates emotional dissonance and emotional exhaus- tion among frontline hotel employees in a developing sub-Saharan country, Nigeria. Specifically, the relationships of negative affectivity as a dispositional personality variable and intrinsic motivation as a personal resource with emotional dissonance and exhaustion are examined. Job performance and turnover inten- tions, which are the two organizationally valued outcomes, are International Journal of Hospitality Management 28 (2009) 349–358 ARTICLE INFO Keywords: Emotional dissonance Emotional exhaustion Hotel employees Nigeria ABSTRACT The purpose of this study was to develop and test a model, which investigated emotional dissonance and emotional exhaustion among employees in frontline service jobs. Data were gathered via self- administered questionnaires from a sample of frontline hotel employees in Nigeria, which is one of the neglected developing sub-Saharan countries in the African continent. The results of the hierarchical multiple regression analyses demonstrated a number of significant direct and partial mediating effects and provided support for the majority of the hypothesized relationships. Specifically, emotional dissonance partially mediated the relationships of negative affectivity and intrinsic motivation with emotional exhaustion. The results also revealed that emotional exhaustion partially mediated the effect of emotional dissonance on turnover intentions. Unexpectedly, emotional dissonance was found to be positively related to job performance. Implications for frontline employees and their managers and directions for future research are discussed. ß 2008 Elsevier Ltd. All rights reserved. * Corresponding author. Tel.: +90 392 630 1116; fax: +90 392 365 1584. E-mail addresses: [email protected] (O.M. Karatepe), [email protected] (K.D. Aleshinloye). 1 Tel.: +90 392 630 1682; fax: +90 392 365 1584. Contents lists available at ScienceDirect International Journal of Hospitality Management journal homepage: www.elsevier.com/locate/ijhosman 0278-4319/$ – see front matter ß 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.ijhm.2008.12.002

Emotional dissonance and emotional exhaustion among hotel employees in Nigeria

  • Upload
    tamu

  • View
    0

  • Download
    0

Embed Size (px)

Citation preview

International Journal of Hospitality Management 28 (2009) 349–358

Emotional dissonance and emotional exhaustion among hotel employees inNigeria

Osman M. Karatepe *, Kayode Dare Aleshinloye 1

School of Tourism and Hospitality Management, Eastern Mediterranean University, Gazimagusa, Turkish Republic of Northern Cyprus, via Mersin 10, Turkey

A R T I C L E I N F O

Keywords:

Emotional dissonance

Emotional exhaustion

Hotel employees

Nigeria

A B S T R A C T

The purpose of this study was to develop and test a model, which investigated emotional dissonance and

emotional exhaustion among employees in frontline service jobs. Data were gathered via self-

administered questionnaires from a sample of frontline hotel employees in Nigeria, which is one of the

neglected developing sub-Saharan countries in the African continent. The results of the hierarchical

multiple regression analyses demonstrated a number of significant direct and partial mediating effects

and provided support for the majority of the hypothesized relationships. Specifically, emotional

dissonance partially mediated the relationships of negative affectivity and intrinsic motivation with

emotional exhaustion. The results also revealed that emotional exhaustion partially mediated the effect

of emotional dissonance on turnover intentions. Unexpectedly, emotional dissonance was found to be

positively related to job performance. Implications for frontline employees and their managers and

directions for future research are discussed.

� 2008 Elsevier Ltd. All rights reserved.

Contents lists available at ScienceDirect

International Journal of Hospitality Management

journa l homepage: www.e lsev ier .com/ locate / i jhosman

1. Introduction

In a turbulent business environment of scarce resources,intense competitive pressures, and rapid rates of technologicalchange, hospitality firms aim to focus on flawless service quality togain and sustain competitive advantages for success and survival.According to the resource-based view, inimitable human resources(e.g., highly qualified and meritorious frontline or customer–contact employees) serve this purpose.

Frontline employees are in frequent face-to-face or voice-to-voice interactions with customers and do emotional labor, which isdefined as ‘‘the effort, planning, and control needed to expressorganizationally desired emotion during interpersonal transac-tions’’ (Morris and Feldman, 1996, p. 987). There are three modes ofacting: surface, active deep, and passive deep (Hochschild, 1983).Surface acting occurs when employees fake their emotions bychanging their outer demeanor to conform with the organizationaldisplay rules while the inner feelings remain unchanged (Zapfet al., 1999). Active deep acting occurs when employees have tospend effort to manage emotions by regulating their expressivebehaviors and inner feelings, and passive deep acting occurs whenemployees may automatically feel emotions needed for a

* Corresponding author. Tel.: +90 392 630 1116; fax: +90 392 365 1584.

E-mail addresses: [email protected] (O.M. Karatepe),

[email protected] (K.D. Aleshinloye).1 Tel.: +90 392 630 1682; fax: +90 392 365 1584.

0278-4319/$ – see front matter � 2008 Elsevier Ltd. All rights reserved.

doi:10.1016/j.ijhm.2008.12.002

particular situation (Zapf et al., 1999). Emotional dissonance,which refers to incongruence between felt and displayed emotions(Heuven and Bakker, 2003), captures Hochschild’s (1983) twomodes of surface acting and passive deep acting, which are theopposite ends of a continuum (Kruml and Geddes, 2000).

Employees in frontline service jobs are susceptible to emotionaldissonance (Phillips et al., 2007). Such employees are alsoconfronted with emotional exhaustion, which is a form of work-related strain and refers to the lack of energy and depletion ofemotional resources due to excessive psychological demands(Boles et al., 2000). Emotional dissonance and/or emotionalexhaustion may give rise to negative job outcomes such as poorjob performance, job dissatisfaction, reduced organizationalcommitment, absenteeism, and turnover (Abraham, 1998; Cro-panzano et al., 2003; Grandey et al., 2004; Karatepe and Uludag,2007; Lewig and Dollard, 2003; Morris and Feldman, 1997).Besides, there are ample tangible and intangible costs associatedwith occupational stress and emotional exhaustion/burnout in aworld of increasing global competition (Gabriel and Liimatainen,2000; Kelloway et al., 2008; Spielberger et al., 2003).

Against this backdrop, this study develops and tests a model,which investigates emotional dissonance and emotional exhaus-tion among frontline hotel employees in a developing sub-Saharancountry, Nigeria. Specifically, the relationships of negativeaffectivity as a dispositional personality variable and intrinsicmotivation as a personal resource with emotional dissonance andexhaustion are examined. Job performance and turnover inten-tions, which are the two organizationally valued outcomes, are

O.M. Karatepe, K.D. Aleshinloye / International Journal of Hospitality Management 28 (2009) 349–358350

treated as the consequences of emotional dissonance andexhaustion. This empirical study treats emotional dissonance asa potential mediator between personality variables and emotionalexhaustion. This study also focuses on emotional exhaustion as apotential mediator between emotional dissonance and job out-comes.

There are potential contributions of this empirical study. First,an observation made from a synthesis of the relevant literatureillustrates that empirical research appertaining to emotionaldissonance and exhaustion in developing countries, especially inthe sub-Saharan African countries, is sparse. Not surprisingly,empirical studies in the services marketing literature using datafrom the developing sub-Saharan African countries are meager(Svensson et al., 2008). Therefore, the present study extends theexisting research stream to Nigeria, which is one of the ignoreddeveloping sub-Saharan African countries. Second, emotional laboris an underresearched topic in the hospitality management andmarketing literatures (Chu and Murrmann, 2006; Kim, 2008). Inaddition, emotional dissonance is the key dimension of emotionallabor pertaining to its impacts on burnout and negative healthoutcomes (Dormann and Zapf, 2004). Hence, this study tests theantecedents and consequences of emotional dissonance amongfrontline hotel employees. Lastly, there will be a variety of usefulmanagement implications based on the findings of this study.

2. Hypotheses and research model

2.1. Hypotheses

The Conservation of Resources (COR) theory is one of thetheories used for developing the relevant relationships in ourstudy, in addition to the perception mechanism and the basictheory of heat-affect-overload. The COR theory defines resourcesas ‘‘. . .those entities that either are centrally valued in their ownright, or act as means to obtain centrally valued ends’’ (Hobfoll,2002, p. 307). According to the theory, there are four fundamentalresource categories: object, personal, condition, and energyresources (Hobfoll, 1989). Individuals seek to acquire, maintainand preserve certain resources. Stress occurs in the workplacewhen (a) individuals are confronted with the threat of loss ofresources, (b) individuals lose their resources, and (c) individualsinvest resources and do not harvest what they have predicted inreturn (Hobfoll, 2001). The two main principles of the COR theoryare (1) ‘‘. . . resource loss is disproportionately more salient thanresource gain’’ (Hobfoll, 2001, p. 343) and (2) ‘‘. . . people mustinvest resources in order to protect against resource loss, recoverfrom losses, and gain resources’’ (Hobfoll, 2001, p. 349). The firstprinciple of the theory has been used to explain the impacts ofstressors on strains, while the second principle of the theory hasbeen used to explain phenomena such as the investment ofpersonal, social, and economic resources in aiding the process ofstress resistance (Grandey and Cropanzano, 1999; Hobfoll, 2001;Ito and Brotheridge, 2003; Yavas et al., 2008). The theory positsthat individuals who are confronted with excessive demands and/or inadequate resources in the workplace may have negativeoutcomes (Shaffer et al., 2001). Additionally, resource loss andburnout are likely to result in certain negative attitudinal andbehavioral consequences (Lee and Ashforth, 1996).

2.1.1. Personality variables, emotional dissonance, and emotional

exhaustion

We examine two personality variables as the antecedents ofemotional dissonance and emotional exhaustion. These antece-dents are negative affectivity and intrinsic motivation. Negativeaffectivity is a dispositional personality variable and an indivi-dual’s tendency to experience discomfort across time and

situations (Watson and Clark, 1984). Individuals high in negativeaffectivity tend to dwell on the negative aspects of themselves,others, and situations in a generally more negative way and oftenseem to be anxious, nervous, and afraid (Cropanzano et al., 1993).Individuals low in negative affectivity are typically in states ofcalmness and serenity (Watson et al., 1988). As discussed by Liuet al. (2004), negative affectivity is likely to arouse negativeemotional experiences across time and situations that may impedeindividuals to regulate their emotional experiences in the serviceencounter. Such individuals appear to fake their positive emotionswhen necessary (Kim, 2008).

The relationship between negative affectivity and stressors isalso supported by the basic theory of heat-affect-overload (Van DeVliert and Van Yperen, 1996). Specifically, employees living andworking in hot climates of countries such as Nigeria, Indonesia, andSingapore are high in negative affectivity and experience roleoverload. It has been proposed that availability of heat or hotclimate disrupts the thermoregulatory system of the human bodyand leads to negative affectivity. Such high negative affectivityindividuals are faced with higher role overload (Van De Vliert andVan Yperen, 1996). Even though the hotels may have functioningair-conditioning systems, regular power cut or outages due to poorelectric power infrastructure in Nigeria may cause frustrationamong employees and customers. In addition, the high cost ofrunning alternative power source like generators limits the properuse of the air-conditioning systems in most of the hotels. Frontlinehotel employees such as door attendants, food servers, andbeverage servers have to serve customers in outdoor facilities.Under these circumstances, such employees are exposed to directsunlight and humidity. Most of the frontline employees cannotafford to buy air-conditioning systems in their houses; if theycould, they do not enjoy it due to irregular power supply in thecountry. Furthermore, they may not have sufficient financialresources to buy automobiles having air-conditioning systems.Therefore, such employees may go to work, usually far from theirhouses, using modes of public transportation such as buses, whichare overloaded and are devoid of air-conditioning systems.Accordingly, frontline hotel employees in a country such asNigeria are high in negative affectivity and experience heightenedstress.

It also has been reported that Nigerians are faced with problemsemerging from job insecurity, long-hours culture, low pay, andgender inequality regarding job opportunities for women, andstressful situations at work (Adeyemi et al., 2006; Aryee, 2005). Anincreasing number of employees in urban sub-Saharan Africa‘‘moonlight as petty traders’’ due to job insecurity and inadequatepay (Aryee, 2005, p. 272). As a result, they are likely to experiencework-family conflict. A very recent empirical study showed thathigh negative affectivity frontline hotel employees in Nigeria werefaced with elevated levels of work-family conflict and family-workconflict (Karatepe and Magaji, 2008). Employees in frontlineservice jobs of the hospitality industry in Nigeria are expected tomanage their emotions by changing their outward appearance todisplay organizationally desired emotions while the inner feelingsremain unchanged and thus are likely to experience emotionalexhaustion.

In addition, negative affectivity is widely used in strain-relatedresearch and has been linked with emotional exhaustion (Houkeset al., 2001). In their meta-analytic work, Thoresen et al. (2003)reported an estimated mean population correlation of .54 betweennegative affectivity and emotional exhaustion. According toSpector et al. (2000), the perception mechanism also proposesuseful guidelines for developing the relevant hypotheses. That is,the perception mechanism states that high negative affectivityindividuals tend to perceive their jobs as stressful and experiencehigh levels of strains. In our case, high negative affectivity frontline

O.M. Karatepe, K.D. Aleshinloye / International Journal of Hospitality Management 28 (2009) 349–358 351

employees in the hotel industry are susceptible to higheremotional dissonance and emotional exhaustion.

Another personality variable used in our study is intrinsicmotivation. To date various personal resources or personalityvariables (e.g., self-efficacy, optimism, and locus of control) havebeen examined with regard to emotional dissonance and emo-tional exhaustion (e.g., Heuven et al., 2006; Ito and Brotheridge,2003; Xanthopoulou et al., 2007). As a personal resource and a keypersonality variable, intrinsic motivation has not received muchempirical attention in the hospitality management and marketingliteratures (cf. Babakus et al., 2008; Karatepe and Uludag, 2007).

Intrinsic motivation refers to an individual’s ‘‘feeling ofchallenge or competence derived from performing a job’’(Keaveney, 1992, p. 151). Intrinsically motivated employees havebetter problem-solving skills and are innovative (Miller, 2002).Grant (2008) states that intrinsically motivated individuals ‘‘feelnaturally drawn, or pulled, toward completing their work’’, ‘‘areprocess focused-they see the work as an end in and of itself’’, and‘‘are present focused-they are concerned with the experience ofperforming the work itself’’ (p. 49). Consistent with the COR theory,intrinsic motivation is one of the personal resources that can beused for coping with emotional dissonance and exhaustion. As apersonal resource, intrinsic motivation can affect employees’willingness and perceived effort to manage emotional experiencesin the service encounter. Such a personal resource can also beinvested in aiding the process of stress resistance and cancontribute to the maintenance of resource reservoirs (cf. Hobfoll,2001). Consequently, employees with personal resources havemastery that enables them to cope with demanding or threateningconditions more effectively and thus prevents them fromexperiencing emotional exhaustion (Xanthopoulou et al., 2007).Recently, Karatepe and Tekinkus (2006) demonstrated thatintrinsic motivation lowered emotional exhaustion for a sampleof frontline bank employees in Turkey. More recently, Karatepeand Uludag (2007) also reported a similar finding among frontlinehotel employees in Northern Cyprus.

The relationship between emotional dissonance and emotionalexhaustion is also developed based on the COR theory, which hasbeen developed to understand the mechanism underlying stress(Halbesleben and Rotondo, 2007). Once employees invest theirscarce resource reservoir, they expect that they will harvest theanticipated rate of return. However, they may not be able to copewith difficulties resulting from emotional dissonance. In addition,they are more sensitive to stressors in the workplace, which threatentheir resources (Shirom, 2003). Employees losing their scantyreservoir in this process are faced with emotional exhaustion.Simply put, emotional dissonance depletes employees’ scarceresources and thus results in emotional exhaustion. There isempirical evidence to support the premise that higher emotionaldissonance leads to intense emotional exhaustion. Broadly speaking,Morris and Feldman (1997) demonstrated a significant positiverelationship between emotional dissonance and emotional exhaus-tion. In a study of cabin attendants in the Netherlands, Heuven andBakker (2003) found that emotional dissonance had a strong positiveimpact on emotional exhaustion. Lewig and Dollard (2003) alsoreported a similar result for a sample of call center employees inAustralia. Recently, it was shown that emotional dissonance wassignificantly and positively related to burnout/emotional exhaus-tion among nurses and police officers in the Netherlands (Bakker andHeuven, 2007). In another study, similar findings were also observedamong Chinese human service employees (Cheung and Tang, 2007).

In light of the aforementioned information, the followinghypotheses are proposed:

H1a. Negative affectivity is positively related to emotional dis-sonance.

H1b. Intrinsic motivation is negatively related to emotional dis-sonance.

H2a. Negative affectivity is positively related to emotionalexhaustion.

H2b. Intrinsic motivation is negatively related to emotionalexhaustion.

H3. Emotional dissonance is positively related to emotionalexhaustion.

In addition, it has been demonstrated that stress (e.g., surfaceacting and workload) partially mediates the effect of negativeaffectivity or neuroticism on strain (Iverson et al., 1998; Kim,2008). The perception mechanism also refers to the mediating roleof stress on the relationship between negative affectivity and strain(Spector et al., 2000). Stress (e.g., surface acting) has been found tohave a partial mediating role on the relationship between personalresources (e.g., core self-evaluations) and strain (Yagil et al., 2008).In line with the COR theory (Hobfoll, 1989), employees withpersonal resources may deal with threatening conditions and thusshield themselves from strain. Accordingly, emotional dissonanceis likely to have a partial mediating role on the relationships ofnegative affectivity and intrinsic motivation with emotionalexhaustion.

Thus, the following hypotheses are proposed:

H4a. Emotional dissonance partially mediates the impact of nega-tive affectivity on emotional exhaustion.

H4b. Emotional dissonance partially mediates the impact ofintrinsic motivation on emotional exhaustion.

2.1.2. Emotional dissonance, emotional exhaustion, and job outcomes

Job performance and turnover intentions are the two joboutcomes examined in this study. Job performance refers to ‘‘thelevel of productivity of an individual employee, relative to his orher peers, on several job-related behaviors and outcomes’’ (Babinand Boles, 1998, p. 82) and turnover intentions refer toemployees’ ‘‘willingness to leave an organization’’ (Thoresenet al., 2003, p. 918). It has been observed that employees infrontline service jobs in the hospitality industry are faced withlow pay, long-hours culture, and inadequate training and work inan environment, which is devoid of family–friendly programsand benefits (Babin and Boles, 1998; Karatepe and Uludag, 2007;Kim et al., 2007; Yavas et al., 2008). Consonant with the CORtheory, employees having excessive demands and insufficientresources in the workplace cannot cope with high levels ofemotional dissonance. As a result, they display poor performanceand turnover intentions.

The COR theory also provides a useful framework fordeveloping the relationships of emotional exhaustion with joboutcomes. Specifically, employees are emotionally exhausted, oncethey find that they lose their scarce resources or face a threat of lossof such resources. Employees cannot perform effectively in theworkplace and demonstrate turnover intentions as a result ofresource loss and emotional exhaustion. In empirical terms,Cropanzano et al. (2003) reported that emotional exhaustioneroded employees’ job performance. In a study conducted withfrontline hotel employees in Turkey, Yavas et al. (2008) found thatemotional exhaustion was significantly and positively associatedwith turnover intentions. However, they found no significantrelationship between emotional exhaustion and job performance.Additionally, Karatepe and Uludag (2007) reported a similarfinding concerning the effect of emotional exhaustion on turnoverintentions.

O.M. Karatepe, K.D. Aleshinloye / International Journal of Hospitality Management 28 (2009) 349–358352

Based on the abovementioned information, the followinghypotheses are proposed:

H5a. Emotional dissonance is negatively related to job perfor-mance.

H5b. Emotional exhaustion is negatively related to job perfor-mance.

H6a. Emotional dissonance is positively related to turnover inten-tions.

H6b. Emotional exhaustion is positively related to turnover inten-tions.

In line with the COR theory (cf. Grandey and Cropanzano, 1999),employees spending their scarce resources for coping withemotional dissonance are faced with emotional exhaustion. Sucha strain may lead to poor job performance. In addition, employeeswho feel emotionally exhausted due to emotional dissonance mayneed to eliminate the resource depletion process by displayingturnover intentions. This discussion suggests that emotionalexhaustion may function as a potential mediator betweenemotional dissonance and job outcomes.

Accordingly, the following hypotheses are proposed:

H7. Emotional exhaustion partially mediates the effects of emo-tional dissonance on (a) job performance and (b) on turnoverintentions.

2.1.3. Control variables

Due to their potential relationships with the relevant studyconstructs, age, gender, education, organizational tenure, andmarital status were controlled for potential confounding influences.For example, research indicated that older service employees wereconfronted with higher emotional dissonance (Kruml and Geddes,2000). In another study, it was shown that younger nurses sufferedfrom emotional exhaustion (Lee and Akhtar, 2007). In contrast, itwas reported that older employees had higher feelings of emotionalexhaustion (Karatepe and Uludag, 2007). Past and recent researchalso linked demographic variables such as age and education withjob outcomes such as job performance, job satisfaction, affectiveorganizational commitment, and turnover intentions (Karatepe andTekinkus, 2006; Martin, 1979).

Fig. 1. Researc

2.2. Research model

The relationships to be tested in the present study are depictedin Fig. 1. Two personality variables are related to emotionaldissonance and emotional exhaustion. That is, negative affectivityhas positive relationships with emotional dissonance and exhaus-tion. However, intrinsic motivation is negatively related toemotional dissonance and exhaustion. The model predicts thatemployees with feelings of emotional dissonance experienceexhaustion. The model also proposes that emotional dissonanceand exhaustion are negatively associated with two organization-ally valued outcomes: job performance and turnover intentions.Broadly, frontline employees experiencing emotional dissonanceand exhaustion display poor job performance and turnoverintentions. As demonstrated in the model, emotional dissonancefunctions as a potential mediator between personality variablesand emotional exhaustion. The model also shows that emotionalexhaustion acts as a potential mediator between emotionaldissonance and job outcomes.

Finally, age, gender, education, organizational tenure, andmarital status are incorporated into the model as control variablesin order to avoid statistical confounds.

3. Method

3.1. Sample and procedure

Data for this empirical study were collected via self-adminis-tered questionnaires in February 2008 from the full-time frontlineemployees of five-star hotels in Abuja and Lagos, Nigeria. Nigeria,which is situated in the sub-Saharan Africa and is on the shores ofGulf of Guinea, gained independence from the Great Britain in 1960and declared itself a Federal Republic in 1963. The country’scomplexity is evident in its ethnicity composition, which includesover 250 ethnic groups, cultural and religious differences, andlanguages. Inhabitants in Nigeria communicate the commonlanguage of English (Adebayo and Udegbe, 2004; Okpara, 2006).Abuja is the capital city, with Lagos being referred to as theeconomic capital city of the country. There are two major seasonsin Nigeria. The dry season is between November and March and therainy season is between April and October. The average daytimetemperature throughout the year in Nigeria is 27 8C, and the

h model.

O.M. Karatepe, K.D. Aleshinloye / International Journal of Hospitality Management 28 (2009) 349–358 353

temperature can go up to 41 8C in the north, while it can reach32 8C in the far south (Nigeria’s First National Communicationunder the UN Framework Convention on Climate Change, 2003).The humidity level in most areas of the country is high. Nigeria isthe most populous black nation in the African continent and is oneof the largest oil exporters in the world (Ituma and Simpson, 2007;Okpara, 2006). Being an oil-rich country, however, Nigeria is stillcharacterized by political instability, corruption, and a high rate ofunemployment (Karatepe and Magaji, 2008).

According to the information obtained from the Director forPlanning, Consultancy and Information Services in the NationalInstitute of Hospitality and Tourism in Abuja at the time of thepresent study, there were seven five-star hotels in Abuja and Lagos,which were licensed by the Nigerian Tourism DevelopmentCorporation. The total number of rooms in these hotels were2734. One of the members of the research team contacted themanagements of seven hotels using a letter outlining theobjectives of this study in Abuja and Lagos. The managementsof five hotels agreed to participate in the study. However, themanagers in these hotels did not permit us to directly contact theirfrontline employees. Therefore, the questionnaires which includedinformation about the assurance of confidentiality were given tothe managers of the hotels. Also the managers were instructed todistribute the questionnaires to a broad range of employees infrontline service positions (e.g., front desk agents, food servers,concierges, bartenders, door attendants, guest relations represen-tatives, and bell attendants). The questionnaires were directlyreturned to one of the members of the research team. Such anapproach is consistent with prior research (Hartline et al., 2000).

120 questionnaires were given to managers in two nationalfive-star hotels in Lagos, while 180 were given to managers of threeinternational five-star chain hotels in Abuja and Lagos. Of the 300questionnaires distributed, a total number of 159 questionnaireswere returned, but 2 questionnaires were eliminated due tomissing information. Consequently, a total number of usable 157questionnaires were retrieved, yielding a response rate of 52.3%.

29% of the respondents were aged between 18 and 27 years, 47%between 28 and 37 years, and the rest were older than 37. Themajority of the respondents (55%) were male. 1% of therespondents had primary school education and 24% secondaryand high school education. 22% of the respondents had 2-yearcollege degrees, 49% 4-year college degrees, and the rest graduatedegrees. 13% of the respondents had tenures of less than 1 year and59% between 1 and 5 years. The rest had tenures more than 5 years.52% of the respondents were married, while the rest were single ordivorced.

3.2. Measurement

Multiple-item scales used in this study were obtained fromdifferent sources in the extant literature.

3.2.1. Negative affectivity

Negative affectivity was measured using three (3) items fromAgho et al. (1992). Sample items include ‘Minor setbacks some-times irritate me too much’, and ‘Often I get irritated at littleannoyances.’ Response options for the negative affectivity itemsranged from 5 (strongly agree) to 1 (strongly disagree). Higher scoresindicated higher negative affectivity.

3.2.2. Intrinsic motivation

Intrinsic motivation was operationalized via four (4) items fromLow et al. (2001). Items include ‘When I do work well, it gives me afeeling of accomplishment’, ‘I feel a great sense of personalsatisfaction when I do my job well’, ‘When I perform my job well, itcontributes to my personal growth and development’, and ‘My job

increases my feeling of self-esteem.’ Such scale items also havebeen used in recent research to operationalize frontline hotel orbank employees’ intrinsic motivation (Babakus et al., 2008;Karatepe and Tekinkus, 2006; Karatepe and Uludag, 2007).Response options for the intrinsic motivation items ranged from5 (strongly agree) to 1 (strongly disagree). Higher scores indicatedhigher intrinsic motivation.

3.2.3. Emotional dissonance

Emotional dissonance was measured using eleven (11) itemsfrom Chu and Murrmann (2006). Sample items include ‘I fake agood mood when interacting with customers’ and ‘I fake theemotions I show when dealing with customers.’ Responses to theitems in emotional dissonance were elicited on five-point scalesranging from 5 (always) to 1 (never). After the positively wordeditems in emotional dissonance had been reverse scored, higherscores indicated higher emotional dissonance.

3.2.4. Emotional exhaustion

Eight (8) items from Maslach and Jackson (1981) were used tomeasure emotional exhaustion. Sample items include ‘I feelemotionally drained from my work’ and ‘I feel used up at theend of the workday.’ The items in emotional exhaustion were ratedon a five-point scale ranging from 5 (strongly agree) to 1 (strongly

disagree). Higher scores indicated higher emotional exhaustion.

3.2.5. Job performance

Five (5) items were adapted from Babin and Boles (1998) tomeasure job performance. Sample items are ‘I am a top performer’and ‘I know what my customers expect better than others.’Response options for the job performance items ranged from 5(strongly agree) to 1 (strongly disagree). Higher scores indicatedhigher job performance.

3.2.6. Turnover intentions

Three (3) items from Singh et al. (1996) were used to measureturnover intentions. Sample items are ‘It is likely that I will activelylook for a new job next year’ and ‘I often think about quitting.’Response options for the turnover intentions items ranged from 5(strongly agree) to 1 (strongly disagree). Higher scores indicatedhigher turnover intentions.

As mentioned before, age, gender, education, organizationaltenure, and marital status were treated as control variables. Age,education, and organizational tenure were measured using five-point scales. Gender was coded as a dichotomous variable(0 = male and 1 = female). Marital status was also coded as adichotomous variable (0 = single or divorced and 1 = married). Inaddition, the questionnaire was tested with a pilot sample of nine(9) frontline employees and there was no compelling reason tomake changes in the questionnaire.

3.3. Data analysis

A confirmatory factor analysis was performed using LISREL 8.30(Joreskog and Sorbom, 1996) to address issues of dimensionality,convergent and discriminant validity (Anderson and Gerbing,1988). The internal consistency reliability was assessed based onthe commonly accepted cut-off value of .70. Hierarchical multipleregression analyses were employed in order to test the relation-ships among the study constructs. The mediation hypotheses weretested using Baron and Kenny (1986) mediation regressiontechnique. In the analysis predicting emotional dissonance, thecontrol variables were entered in step 1 and personality variablesin step 2. The control variables were entered in step 1, personalityvariables in step 2, and emotional dissonance in step 3 in order topredict emotional exhaustion. The same procedure was also used

O.M. Karatepe, K.D. Aleshinloye / International Journal of Hospitality Management 28 (2009) 349–358354

for predicting job outcomes. Specifically, the control variables wereentered in step 1, emotional dissonance in step 2, and emotionalexhaustion in step 3. The same procedure was repeated forturnover intentions. Consistent with Van Dijk and Kirk (2006)study, the Aroian version of the Sobel test was used to examine thesignificance of the mediating roles of emotional dissonance andexhaustion.

4. Results

4.1. Measurement results

Several items such as one item from intrinsic motivation andfour items from emotional dissonance were dropped due to non-significant t-values and low standardized factor loadings (<.40)during confirmatory factor analysis. Specifically, one item ‘My jobincreases my feeling of self-esteem’ from the intrinsic motivationscale was dropped during confirmatory factor analysis. Four items‘The emotions I show to customers match what I truly feel’, ‘I puton an act in order to deal with customers in an appropriate way’, ‘Iactually feel the emotions that I need to show to do my job well’,and ‘I show the same feelings to customers that I feel inside’ fromthe emotional dissonance scale were removed from furtheranalysis. Accordingly, the results yielded the following fitstatistics: x2 = 841.39, df = 362; GFI [goodness of fit index] = .73;AGFI [adjusted goodness of fit index] = .67; NFI [normed fitindex] = .72; NNFI [non-normed fit index] = .78; CFI [comparativefit index]=.81; RMSEA [root mean square error of approxima-tion] = .092; SRMR [standardized root mean square resi-dual] = .078. The x2/df value of 2.32 was within a range ofacceptable values (2–5) (Marsh and Hocevar, 1985) and the resultfor SRMR fell within a range of acceptable value (<.10) (Sumeret al., 2000). The results for GFI, AGFI, NFI, NNFI, and CFI were lowerthan .90 (Bentler and Bonett, 1980) and the result for RMSEA washigher than .08 (Browne and Cudeck, 1993). Overall, the fitstatistics were unsatisfactory. However, the magnitudes of theloadings ranged from .46 to .92 and their t-values (>2.00) weresignificant. Of the 29 items, 19 had loadings greater than .70.Consequently, the results provided evidence of convergent validity(Anderson and Gerbing, 1988).

Discriminant validity was evaluated based on a series of x2

difference tests using measures of each pair of constructs. Such anapproach is consistent with the other empirical studies (Babakuset al., 1996; Low et al., 2001). In particular, a two-dimensionalmodel for each pair of constructs was first fit, and then itemsrepresenting each construct were forced into a single-factorsolution. As depicted in Table 1, the x2 difference test produced asignificant result for each pair of measures. Thus, imposing a singlefactor solution on the two sets of items representing differentconstructs demonstrated a significant deterioration of the modelfit. These results provided evidence of discriminant validity(Anderson and Gerbing, 1988).

Consonant with the other recent empirical studies (e.g.,Brockman and Morgan, 2006; Kandemir et al., 2006), common

Table 1Discriminant validity: chi-square difference (n = 157).

Variables 1 2 3 4 5 6

1. NA (negative affectivity)

2. INTMOT (intrinsic motivation) 115.57

3. ED (emotional dissonance) 106.57 90.52

4. EEXHAUST (emotional exhaustion) 44.48 77.93 566.14

5. JPERF (job performance) 285.39 86.31 258.94 295.77

6. TINT (turnover intentions) 116.63 72.60 254.41 227.24 260.76

Notes: Chi-square difference tests (all tests = 1 df).

x2 > 10.83, p < .001.

method bias was checked with a confirmatory factor analysisapproach to Harman’s single-factor test as a statistical remedy.This test is based on the assumption that common method bias is aserious problem when a single latent factor will account for morethan 50% of the total variance of the measures (Podsakoff et al.,2003). The results for a single-factor model were as follows: x2

=2078.41, df = 377; GFI = .52; AGFI = .45; NFI = .45; NNFI = .47;CFI = .50; RMSEA = .17; SRMR = .13. The single-factor modelaccounted for only 31% of the total variance. The chi-square testalso demonstrated that the six-factor model was superior to thesingle-factor model (Dx2 = 1237.02, Ddf = 15, p < .001). Conse-quently, the results were worse than that of a six-factor model andone single-factor model did not account for the majority of thevariance. In short, common method bias was not a critical threat tothe magnitudes of the hypothesized relationships.

Also consistent with the suggestion made by Armstrong andOverton (1977) and past research (e.g., Harris and Goode, 2004),non-response bias was controlled through independent t-tests(p < .01). The results yielded no significant differences betweenearly (75%) and late (25%) respondents in terms of the studyvariables. Accordingly, the results demonstrated that non-response did not appear to be a problem in the current study.

Composite scores for each construct were created by averagingall items comprising that particular construct. The results ofmeans, standard deviations, and correlations among the variablesare shown in Table 2. The results revealed that older employeeswere low in negative affectivity, experienced less emotionaldissonance and emotional exhaustion, and had lower turnoverintentions. The results further demonstrated that female employ-ees were faced with higher emotional exhaustion, whereas bettereducated employees were confronted with less emotionaldissonance and exhaustion. The findings in Table 2 indicated thatmarried employees and the ones with longer tenure were low innegative affectivity and were susceptible to low levels of emotionaldissonance and exhaustion, and had less turnover intentions. Asdemonstrated in Table 2, the scales’ alpha reliabilities (Cronbach’salpha) were as follows: negative affectivity .79; intrinsic motiva-tion .69; emotional dissonance .89; emotional exhaustion .92; jobperformance .79; and turnover intentions .90. All coefficient alphasexceeded the .70 benchmark, excluding intrinsic motivation whosecoefficient alpha was slightly below .70.

4.2. Model test results

Table 3 demonstrates the results of the hierarchical multipleregression analyses regarding the direct effects of control variablesand personality variables on emotional dissonance. The resultsdemonstrated that negative affectivity was positively associatedwith emotional dissonance (b = .41, p < .001), while intrinsicmotivation was negatively associated with emotional dissonance(b = �.20, p < .01). Therefore, Hypotheses 1a and 1b weresupported. Of the control variables, only education was found tobe significantly associated with emotional dissonance (b = �.29,p < .001). That is, more educated employees were faced with lessemotional dissonance. The control variables and personalityvariables explained 38% of the variance in emotional dissonance.

According to the results in Table 3, negative affectivity had apositive relationship with emotional exhaustion (b = .61, p < .001)and intrinsic motivation was negatively related to emotionalexhaustion (b = �.17, p < .01). Thus, Hypotheses 2a and 2b weresupported. The results also showed that emotional dissonance waspositively related to emotional exhaustion (b = .21, p < .01).Therefore, Hypothesis 3 was supported. It was observed inTable 3 that when emotional dissonance was entered in theequation, the sizes of the direct effects of negative affectivity andintrinsic motivation on emotional exhaustion were reduced, but

Table 2Means, standard deviations, correlations of study variables, and Cronbach’s alpha.

Variables # of items Alpha 1 2 3 4 5 6 7 8 9 10 11

1. Age – – 1.000

2. Gender – – �.063 1.000

3. Education – – .216** .063 1.000

4. Organizational tenure – – .654** �.019 .179* 1.000

5. Marital status – – .548** �.105 .181* .448** 1.000

6. Negative affectivity (NA) 3 .79 �.226** .129 �.153 �.167* �.211** 1.000

7. Intrinsic motivation (INTMOT) 4a .69 .110 �.037 .133 .083 .091 �.094 1.000

8. Emotional dissonance (ED) 11b .89 �.175* .152 �.371** �.170* �.195* .484** �.278** 1.000

9. Emotional exhaustion (EEXHAUST) 8 .92 �.221** .174* �.228** �.184* �.174* .664** �.251** .538** 1.000

10. Job performance (JPERF) 5 .79 �.104 �.007 �.129 �.073 �.086 .168* .157* .238** .043 1.000

11. Turnover intentions (TINT) 3 .90 �.263** .134 �.152 �.253** �.177* .418** �.301** .523** .575** .159* 1.000

Mean – – 2.00 .45 3.33 2.24 .52 3.04 4.60 3.19 2.81 3.98 3.42

Standard deviation – – .82 .50 .91 .80 .50 1.01 .48 .99 .91 .65 1.08

Notes: Composite scores for each measure were obtained by averaging scores across items representing that measure. The scores range from 1 to 5. Age, education, and

organizational tenure were measured using five-point scales. Higher scores indicated older age, more educated, and longer tenure. Gender was coded as a dichotomous

variable (0 = male and 1 = female). Marital status was also coded as a dichotomous variable (0 = single or divorced and 1 = married).a One item from intrinsic motivation was dropped during confirmatory factor analysis.b Four items were dropped from emotional dissonance during confirmatory factor analysis.* Correlations are significant at the .05 level.** Correlations are significant at the .01 level. Correlations without any asterisks are not significant.

O.M. Karatepe, K.D. Aleshinloye / International Journal of Hospitality Management 28 (2009) 349–358 355

remained significant. There was a significant increment in R2 of themodel (DR2 = .03, p < .01). Such findings indicated that emotionaldissonance partially mediated the impacts of negative affectivityand intrinsic motivation on emotional exhaustion.

As mentioned before, the Aroian version of the Sobel test wasconducted for establishing the significance of mediating role ofemotional dissonance on the relationship between personalityvariables and emotional exhaustion. In Table 3, the resultsdemonstrated that emotional dissonance partially mediated therelationship between negative affectivity and emotional exhaustion

Table 3Hierarchical multiple regression results: direct and mediating effects.

Independent variables Standardized regression weights

Emotional disso-

nance

Emotional exhaustion

Step 1 Step 2 Step 1 Step 2 Step 3

(I) Control variables

Age .00 .07 �.11 �.02 �.04

Gender .16* .11 .18* .10 .07

Education �.35*** �.29*** �.20* �.11 �.05

Organizational tenure �.07 �.06 �.06 �.05 �.04

Marital status �.09 �.04 �.03 .03 .04

(II) Personality variables

Negative affectivity .41*** .61*** .53***

Intrinsic motivation �.20** �.17** �.13*

(III) Emotional dissonance – – .21**

F 6.81*** 13.02*** 4.00** 21.27*** 20.69***

R2 at each step .18 .38 .12 .50 .53

DR2 – .20*** – .38*** .03**

Aroian test for

NA! ED! EEXHAUST 2.63**

INTMOT! ED! EEXHAUST �2.06*

Note: Age, education, and organizational tenure were measured using five-point

scales. Higher scores indicated older age, more educated, and longer tenure. Gender

was coded as a dichotomous variable (0 = male and 1 = female). Marital status was

also coded as a dichotomous variable (0 = single or divorced and 1 = married).

NA = negative affectivity; INTMOT = intrinsic motivation; ED = emotional disso-

nance; EEXHAUST = emotional exhaustion. The results regarding variance inflation

factors were below 2.1 and did not demonstrate any problems of multicollinearity.* p < .05.** p < .01.*** p < .001.

(test statistic = 2.63, p < .01) and the relationship between intrinsicmotivation and emotional exhaustion (test statistic =�2.06,p < .05). Therefore, Hypotheses 4a and 4b were supported. Genderand education demonstrated significant relationships with emo-tional exhaustion in step 1. However, they did not significantlyimpact emotional exhaustion due to the inclusion of the personalityvariables and/or emotional dissonance in the equation. The controlvariables, personality variables, and emotional dissonanceexplained 53% of the variance in emotional exhaustion.

The results in Table 4 demonstrated that emotional dissonancepositively affected job performance in both steps 2 (b = .22,p < .05) and 3 (b = .29, p < .01). These results led to the conclusionthat Hypothesis 5a was not supported. Emotional exhaustion wasnot found to be significantly related to job performance. Thus,hypothesis 5b was not supported. Such findings also demonstratedthat Hypothesis 7a, which referred to the mediating role ofemotional exhaustion on the relationship between emotionaldissonance and job performance, was not supported. None of thecontrol variables was significantly associated with job perfor-mance. The control variables, emotional dissonance, and emotionalexhaustion explained 8% of the variance in job performance.

It was observed in Table 4 that emotional dissonance positivelyimpacted turnover intentions (b = .51, p < .001). Therefore,Hypothesis 6a was supported. The results indicated that emotionalexhaustion positively influenced turnover intentions (b = .39,p < .001). Hence, Hypothesis 6b was supported. When emotionalexhaustion was entered in the model, the size of the direct impactof emotional dissonance on turnover intentions was reduced, butremained significant. The results demonstrated a significantincrement in R2 of the model (DR2 = .10, p < .001). Accordingly,emotional exhaustion was found to partially mediate the effect ofemotional dissonance on turnover intentions.

The Aroian version of the Sobel test was also conducted forestablishing the significance of mediating role of emotionalexhaustion on the relationship between emotional dissonanceand turnover intentions. The results in Table 4 showed asignificant mediating role of emotional exhaustion on therelationship between emotional dissonance and turnover inten-tions (test statistic = 2.54, p < .05). Thus, Hypothesis 7b wassupported. None of the control variables had significant relation-ships with turnover intentions. The control variables, emotionaldissonance, and emotional exhaustion explained 42% of thevariance in turnover intentions.

Table 4Hierarchical multiple regression results: direct and mediating effects.

Independent variables Standardized regression weights

Job performance Turnover intentions

Step 1 Step 2 Step 3 Step 1 Step 2 Step 3

(I) Control variables

Age �.07 �.07 �.08 �.14 �.14 �.10

Gender �.01 �.04 �.03 .13 .05 .01

Education �.11 �.03 �.03 �.10 .08 .09

Organizational tenure .00 .02 .02 �.14 �.11 �.10

Marital status �.03 �.01 �.01 �.01 .04 .03

(II) Emotional dissonance .22* .29** .51*** .32***

(III) Emotional exhaustion �.13 .39***

F .73 1.70 1.73 3.58** 11.65*** 15.53***

R2 at each step .02 .06 .08 .11 .32 .42

DR2 – .04* .02 – .21*** .10***

Aroian test for

ED! EEXHAUST! JPERF na

ED! EEXHAUST! TINT 2.54*

Note: Age, education, and organizational tenure were measured using five-point scales. Higher scores indicated older age, more educated, and longer tenure. Gender was

coded as a dichotomous variable (0 = male and 1 = female). Marital status was also coded as a dichotomous variable (0 = single or divorced and 1 = married). ED = emotional

dissonance; EEXHAUST = emotional exhaustion; JPERF = job performance; TINT = turnover intentions; na = not applicable. The results regarding variance inflation factors

were below 2.1 and did not demonstrate any problems of multicollinearity.* p < .05.** p < .01.*** p < .001.

O.M. Karatepe, K.D. Aleshinloye / International Journal of Hospitality Management 28 (2009) 349–358356

5. Discussion

The focus of this study was on developing and testing a model,which investigated emotional dissonance and emotional exhaus-tion among frontline hotel employees in Nigeria. The current studyextended this research stream to Nigeria, which is one of theunderrepresented developing sub-Saharan African countries. Theantecedents of emotional dissonance and exhaustion examined inthe present study were negative affectivity and intrinsic motiva-tion. Their consequences investigated here were job performanceand turnover intentions. In the current study, emotional dis-sonance and emotional exhaustion were also treated as potentialmediators. Of the 13 hypotheses, 10 were supported and a numberof useful findings emerged from this empirical investigation.

5.1. Assessment of findings

The results demonstrating that negative affectivity is positivelyrelated to emotional dissonance and exhaustion are concordantwith the perception mechanism (Spector et al., 2000). The resultsthat intrinsic motivation depicts a negative relationship withemotional dissonance and exhaustion are in line with the CORtheory (Hobfoll, 1989, 2001). The result pertaining to therelationship of intrinsic motivation with emotional exhaustion isalso consonant with those of Babakus et al. (2008) and Karatepeand Uludag (2007). As hypothesized, emotional dissonance ispositively related to frontline employees’ emotional exhaustion.This is congruent with the COR theory (Hobfoll, 1989, 2001) andother empirical studies (Heuven and Bakker, 2003; Morris andFeldman, 1997). Although service with a smile is a prevalent jobrequirement during the selection of new employees in servicefirms (Goldberg and Grandey, 2007), employees experiencingemotional dissonance on a continuing basis lose their scantyresource reservoir and are faced with emotional exhaustion, whichgives rise to tangible and intangible costs for organizations.

The result, which pertains to the partial mediating role ofemotional dissonance on the relationship between negativeaffectivity and emotional exhaustion is also consistent with the

perception mechanism. Broadly, frontline employees tending toperceive themselves, others, and events in a generally more negativemanner are faced with emotional dissonance and thus exhaustion.The results of this study also demonstrate that emotional dissonancepartially mediates the effect of intrinsic motivation on emotionalexhaustion. Consistent with the COR theory (Hobfoll, 1989),employees with intrinsic motivation may cope with emotionaldissonance and thus protect themselves from emotional exhaustion.

Surprisingly, it has been observed that emotional dissonance hasa positive effect on frontline employees’ job performance. A veryrecent study also has demonstrated a similar finding concerning theimpact of a stressor on job performance. Specifically, it has beenillustrated that frontline hotel employees who are confronted withconflict between work and family domains display effectiveperformance in the workplace (Yavas et al., 2008). The interestingfinding reported in the current study can be explained based on VanDyne et al. (2002), Singh et al. (1994), Ashforth and Humphrey(1993), and Yavas et al. (2008) studies. First, employees who aresusceptible to heightened emotional dissonance try to concentrateon their job-related duties and responsibilities more to protectthemselves from experiencing further emotional dissonance and tobe able to perform effectively in the organization (Van Dyne et al.,2002; Yavas et al., 2008). Second, two countervailing effects such asthe eustress-type (positive/functional) and the distress-type (nega-tive/functional) can be used for explaining the relationship betweenemotional dissonance and job performance (Singh et al., 1994; Yavaset al., 2008). As discussed by Yavas et al. (2008), such countervailingeffects may act simultaneously to lessen the effect of a stressor on jobperformance. In our case, the eustress-type appears to be the moredominant component of such an effect on job performance. Third, indynamic service encounters, emotional dissonance might haveprovided frontline employees with a critical means of regulatingtheir interaction with customers. This might have resulted in taskeffectiveness (Ashforth and Humphrey, 1993). Whatever theexplanation for this unexpected finding is, the relationship betweenthe two constructs needs further empirical attention. On the otherhand, emotional exhaustion does not have any bearing on jobperformance. The use of self-report measure of job performance may

O.M. Karatepe, K.D. Aleshinloye / International Journal of Hospitality Management 28 (2009) 349–358 357

be responsible for such an insignificant finding. Consequently,emotional exhaustion does not act as a mediator between emotionaldissonance and job performance.

The results suggest that frontline employees display higherturnover intentions due to elevated levels of emotional dissonanceand exhaustion and emotional exhaustion functions as a partialmediator between emotional dissonance and turnover intentions.Such findings are in accordance with the COR theory (cf. Grandeyand Cropanzano, 1999).

5.2. Practical implications

As reported in previous studies, poor human resource manage-ment practices are evident in the hospitality industry (Guerrier andDeery, 1998; Karatepe and Uludag, 2007). Such a reality is also viablefor the hospitality industry in Nigeria. As potential remedies, theresults of this study suggest important implications for managerialaction.Oneofthemisthatmanagersshoulduseeffectiverecruitmentand selection tools in order to fill vacant positions in theirorganizations. It is significant that managers should consider thepersonality traits of candidates in the process, focus on candidateswho are intrinsically motivated, and try to hire those who canmanage their felt emotions matching organizationally desireddisplay norms in the service encounter. This should be a far-reachingpractice among hospitality managers. By doing so, managers can gosome way to making such frontline employees manage problemsassociated with emotional dissonance and exhaustion. Anotherimplication for practice is that employing mentors in the workplaceappears to be inevitable, since younger, less educated and lessexperienced employees are confronted with emotional dissonanceand exhaustion. Mentors could help such employees alleviate theiremotional dissonance and exhaustion by listening to employees’problems and their expectations from the management of the hoteland providing support and guidance (cf. Lee and Akhtar, 2007).

In addition, frontline employees should be trained continuouslyto learn how to cope with problems that stem from emotionaldissonance and emotional exhaustion. This is significant, becauseeffective and continuous training programs in the hospitalityindustry are not abundant. Therefore, managers should foster socialsupport emanating from both supervisors and coworkers in theworkplace during these training programs and train their frontlineemployees in the areas of complaint handling procedures andgenuine customer care. Such training programs would also consist ofpotential empowerment practices frontline employees would use todeal with customers’ complaints. The final implication is associatedwith promotional opportunities in the workplace. Managers shouldheed promoting frontline employees to supervisory positions whocan cope successfully with stressors and strains and make significantcontributions to teamwork. Such successful employees may serve asrole models and create a demonstration effect among the otheremployees who are high in negative affectivity.

6. Concluding remarks

Although this study delineates useful findings and implicationsfor managers and employees, like all empirical studies, it hasseveral limitations. One of the limitations is that the current studyused cross-sectional data to test the relationships in the conceptualmodel. Although previous and recent theoretical and empiricalstudies provide support and/or useful guidelines for the directionof the relationships among the study variables, the cross-sectionalnature of the data does not permit a true test of causality. With thisrecognition, future studies should use longitudinal data to confirmthe findings reported in this study. Another limitation is associatedwith the potential threat of common method bias. A confirmatoryfactor analysis approach to Harman’s single-factor test was used in

order to check such a potential threat. The results indicated thatcommon method bias did not pose a threat. However, it would bemuch more reliable and beneficial to collect data from differentsources (managers and/or customers) in future research to avoidthis problem. Moreover, the current study did not use heat ortemperature as an independent variable to test its relationshipwith negative affectivity and emotional dissonance. In futureresearch, including it in the model would be more fruitful. The finallimitation pertains to the use of one single industry of one singlecountry. Therefore, replication studies with larger sample sizes indifferent service settings (e.g., travel agencies and airlines) of theother developing sub-Saharan African countries would be bene-ficial to broaden the database for further generalizations. Inclosing, despite these limitations, the findings of this study haveshed light on the understanding of emotional dissonance andexhaustion among frontline hotel employees in one of theneglected developing sub-Saharan African countries, Nigeria.

References

Abraham, R., 1998. Emotional dissonance in organizations: antecedents, conse-quences, and moderators. Genetic, Social and General Psychology Monographs124 (2), 229–246.

Adebayo, D.O., Udegbe, I.B., 2004. Gender in the boss–subordinate relationship: aNigerian study. Journal of Organizational Behavior 25 (4), 515–525.

Adeyemi, A.Y., Ojo, S.O., Aina, O.O., Olanipekun, E.A., 2006. Empirical evidence ofwomen under-representation in the construction industry in Nigeria. Womenin Management Review 21 (7), 567–577.

Agho, A.O., Price, J.L., Mueller, C.W., 1992. Discriminant validity of measures of jobsatisfaction, positive affectivity and negative affectivity. Journal of Occupationaland Organizational Psychology 65 (3), 185–196.

Anderson, J.C., Gerbing, D.W., 1988. Structural equation modeling in practice: areview and recommended two-step approach. Psychological Bulletin 103(May), 411–423.

Armstrong, J.S., Overton, T.S., 1977. Estimating nonresponse bias in mail surveys.Journal of Marketing Research 14 (August), 396–402.

Aryee, S., 2005. The work-family interface in urban sub-Saharan Africa: a theoreticalanalysis. In: Poelmans, S.A. (Ed.), Work and Family: An International ResearchPerspective. Lawrence Erlbaum Associates, Incorporated, Mahwah, New Jersey,NJ, pp. 261–286.

Ashforth, B.E., Humphrey, R.H., 1993. Emotional labor in service roles: the influenceof identity. Academy of Management Review 18 (1), 88–115.

Babakus, E., Cravens, D.W., Johnston, M., Moncrief, W.C., 1996. Examining the role oforganizational variables in the salesperson satisfaction model. Journal of Per-sonal Selling and Sales Management 16 (3), 33–46.

Babakus, E., Yavas, U., Karatepe, O.M., 2008. The effects of job demands, jobresources and intrinsic motivation on emotional exhaustion and turnoverintentions: a study in the Turkish hotel industry. International Journal ofHospitality and Tourism Administration 9 (4), 384–404.

Babin, B., Boles, J.S., 1998. Employee behavior in a service environment: a model andtest of potential differences between men and women. Journal of Marketing 62(April), 77–91.

Bakker, A.B., Heuven, E., 2007. Emotional dissonance, burnout, and in-role perfor-mance among nurses and police officers. International Journal of Stress Man-agement 13 (4), 423–440.

Baron, R.M., Kenny, D.A., 1986. The moderator–mediator variable distinction insocial psychological research: conceptual, strategic, and statistical considera-tions. Journal of Personality and Social Psychology 51 (6), 1173–1182.

Bentler, P.M., Bonett, D.G., 1980. Significance tests and goodness of fit in the analysisof covariance structures. Psychological Bulletin 88 (3), 588–606.

Boles, J.S., Dean, D.H., Ricks, J.M., Short, J.C., Wang, G., 2000. The dimensionality ofthe Maslach Burnout Inventory across small business owners and educators.Journal of Vocational Behavior 56 (1), 12–34.

Brockman, B.K., Morgan, R.M., 2006. The moderating effect of organizationalcohesiveness in knowledge use and new product development. Journal ofthe Academy of Marketing Science 34 (3), 295–307.

Browne, M.W., Cudeck, R., 1993. Alternative ways of assessing model fit. In: Bollen,K.A., Long, J.S. (Eds.), Testing Structural Equation Models. Sage, Newbury Park,CA, pp. 136–162.

Cheung, F.Y.-L., Tang, C.S.-K., 2007. The influence of emotional dissonance andresources at work on job burnout among Chinese human service employees.International Journal of Stress Management 14 (1), 72–87.

Chu, K.H-L., Murrmann, S.K., 2006. Development and validation of the hospitalityemotional labor scale. Tourism Management 27 (6), 1181–1191.

Cropanzano, R., James, K., Konovsky, M.A., 1993. Dispositional affectivity as apredictor of work attitudes and job performance. Journal of OrganizationalBehavior 14 (6), 595–606.

Cropanzano, R., Rupp, D.E., Byrne, Z.S., 2003. The relationship of emotional exhaus-tion to work attitudes, job performance, and organizational citizenship beha-viors. Journal of Applied Psychology 88 (1), 160–169.

O.M. Karatepe, K.D. Aleshinloye / International Journal of Hospitality Management 28 (2009) 349–358358

Dormann, C., Zapf, D., 2004. Customer-related social stressors and burnout. Journalof Occupational Health Psychology 9 (1), 61–82.

Gabriel, P., Liimatainen, M.-R., 2000. Mental Health in the Workplace. InternationalLabor Office, Geneva, Switzerland.

Goldberg, L.S., Grandey, A.A., 2007. Display rules versus display autonomy: emotionregulation, emotional exhaustion, and task performance in a call center simula-tion. Journal of Occupational Health Psychology 12 (3), 301–318.

Grandey, A.A., Cropanzano, R., 1999. The conservation of resources model applied towork-family conflict and strain. Journal of Vocational Behavior 54 (2), 350–370.

Grandey, A.A., Dickter, D.N., Sin, H.-P., 2004. The customer is not always right:customer aggression and emotion regulation of service employees. Journal ofOrganizational Behavior 25 (3), 397–418.

Grant, A.M., 2008. Does intrinsic motivation fuel the prosocial fire? Motivationalsynergy in predicting persistence, performance, and productivity. Journal ofApplied Psychology 93 (1), 48–58.

Guerrier, Y., Deery, M., 1998. Research in hospitality human resource managementand organizational behavior. International Journal of Hospitality Management17 (2), 145–160.

Halbesleben, J.R.B., Rotondo, D.M., 2007. Developing social support in employees:human resource development lessons from same-career couples. Advances inDeveloping Human Resources 9 (4), 544–555.

Harris, L.C., Goode, M.M.H., 2004. The four levels of loyalty and the pivotal role oftrust: a study of online service dynamics. Journal of Retailing 80 (2), 139–158.

Hartline, M.D., Maxham, J.G., McKee, D.O., 2000. Corridors of influence in thedissemination of customer-oriented strategy to customer contact serviceemployees. Journal of Marketing 64 (April), 35–50.

Heuven, E., Bakker, A.B., 2003. Emotional dissonance and burnout among cabin atten-dants. European Journal of Work and Organizational Psychology 12 (1), 81–100.

Heuven, E., Bakker, A.A., Schaufeli, W.B., Huisman, N., 2006. The role of self-efficacyin performing emotion work. Journal of Vocational Behavior 69 (2), 222–235.

Hobfoll, S.E., 1989. Conservation of resources: a new attempt at conceptualizingstress. American Psychologist 44 (3), 513–524.

Hobfoll, S.E., 2001. The influence of culture, community, and the nested-self in thestress process: advancing conservation of resources theory. Applied Psychol-ogy: An International Review 50 (3), 337–421.

Hobfoll, S.E., 2002. Social and psychological resources and adaptation. Review ofGeneral Psychology 6 (4), 307–324.

Hochschild, A.R., 1983. The Managed Heart: Commercialization of Human Feeling.University of California Press, Berkeley, CA.

Houkes, I., Janssen, P.P.M., De Jonge, J., Nijhuis, F.J.N., 2001. Work and individualdeterminants of intrinsic work motivation, emotional exhaustion, and turnoverintention: a multi-sample analysis. International Journal of Stress Management8 (4), 257–283.

Ito, J.K., Brotheridge, C.M., 2003. Resources, coping strategies, and emotionalexhaustion: a conservation of resources perspective. Journal of VocationalBehavior 63 (3), 490–509.

Ituma, A., Simpson, R., 2007. Moving beyond Schein’s typology: individual careeranchors in the context of Nigeria. Personnel Review 36 (6), 978–995.

Iverson, R.D., Olekalns, M., Erwin, P.J., 1998. Affectivity, organizational stressors, andabsenteeism: a causal model of burnout and its consequences. Journal ofVocational Behavior 52 (1), 1–23.

Joreskog, K., Sorbom, D., 1996. LISREL 8: User’s Reference Guide. Scientific SoftwareInternational, Inc., Chicago.

Kandemir, D., Yaprak, A., Cavusgil, S.T., 2006. Alliance orientation: conceptualiza-tion, measurement, and impact on market performance. Journal of the Academyof Marketing Science 34 (3), 324–340.

Karatepe, O.M., Magaji, A.B., 2008. Work-family conflict and facilitation in the hotelindustry: a study in Nigeria. Cornell Hospitality Quarterly 49 (4), 395–412.

Karatepe, O.M., Tekinkus, M., 2006. The effects of work-family conflict, emotionalexhaustion, and intrinsic motivation on job outcomes of frontline employees.International Journal of Bank Marketing 24 (3), 173–193.

Karatepe, O.M., Uludag, O., 2007. Conflict, exhaustion, and motivation: a study offrontline employees in Northern Cyprus hotels. International Journal of Hospi-tality Management 26 (3), 645–665.

Keaveney, S.M., 1992. An empirical investigation of dysfunctional organizationalturnover among chain and non-chain retail store buyers. Journal of Retailing 68(2), 145–173.

Kelloway, E.K., Teed, M., Kelley, E., 2008. The psychosocial environment: towards anagenda for research. International Journal of Workplace Health Management 1(1), 50–64.

Kim, H.J., 2008. Hotel service providers’ emotional labor: the antecedents andeffects on burnout. International Journal of Hospitality Management 27 (2),151–161.

Kim, H.J., Shin, K.H., Umbreit, W.T., 2007. Hotel job burnout: the role of personalitycharacteristics. International Journal of Hospitality Management 26 (2), 421–434.

Kruml, S.M., Geddes, D., 2000. Exploring the dimensions of emotional labor.Management Communication Quarterly 14 (1), 8–49.

Lee, J.S.Y., Akhtar, S., 2007. Job burnout among nurses in Hong Kong: implicationsfor human resource practices and interventions. Asian Pacific Journal of HumanResources 45 (1), 63–84.

Lee, R.T., Ashforth, B.E., 1996. A meta-analytic examination of the correlates of thethree dimensions of job burnout. Journal of Applied Psychology 81 (2), 123–133.

Lewig, K.A., Dollard, M.F., 2003. Emotional dissonance, emotional exhaustion andjob satisfaction in call center workers. European Journal of Work and Organiza-tional Psychology 12 (4), 366–392.

Liu, Y., Perrewe, P.L., Hochwarter, W.A., Kacmar, C.J., 2004. Dispositional antece-dents and consequences of emotional labor at work. Journal of Leadership andOrganizational Studies 10 (4), 12–25.

Low, G.S., Cravens, D.W., Grant, K., Moncrief, W.C., 2001. Antecedents and con-sequences of salesperson burnout. European Journal of Marketing 35 (5/6),587–611.

Marsh, H.W., Hocevar, D., 1985. Application of confirmatory factor analysis to thestudy of self-concept: first- and higher order factor models and their invarianceacross groups. Psychological Bulletin 97 (3), 562–582.

Martin, T.N., 1979. A contextual model of employee turnover intentions. Academyof Management Journal 22 (2), 313–324.

Maslach, C., Jackson, S.E., 1981. The measurement of experienced burnout. Journalof Occupational Behavior 2 (2), 99–113.

Miller, J.F., 2002. Motivating people. Executive Excellence 19 (December), 12.Morris, J.A., Feldman, D.C., 1996. The dimensions, antecedents, and consequences of

emotional labor. Academy of Management Review 21 (4), 986–1010.Morris, J.A., Feldman, D.C., 1997. Managing emotions in the workplace. Journal of

Managerial Issues 9 (3), 257–274.Nigeria’s First National Communication under the UN Framework Convention on

Climate Change, 2003. The Ministry of Environment of the Federal Republic ofNigeria, Abuja (November).

Okpara, J.O., 2006. Gender and the relationship between perceived fairness in pay,promotion, and job satisfaction in a sub-Saharan African economy. Women inManagement Review 21 (3), 224–240.

Phillips, B., Tan, T.T.W., Julian, C., 2007. The theoretical underpinnings of emotionaldissonance: a framework and analysis of propositions. Journal of ServicesMarketing 20 (7), 471–478.

Podsakoff, P.M., MacKenzie, S.B., Lee, J.-Y., Podsakoff, N.P., 2003. Common methodbiases in behavioral research: a critical review of the literature and recom-mended remedies. Journal of Applied Psychology 88 (5), 879–903.

Shaffer, M.A., Harrison, D.A., Gilley, K.M., Luk, D.M., 2001. Struggling for balanceamid turbulence on international assignments: work-family conflict, supportand commitment. Journal of Management 27 (1), 99–121.

Shirom, A., 2003. Job-related burnout. In: Quick, J.C., Tetrick, L.E. (Eds.), Handbookof Occupational Health Psychology. American Psychological Association,Washington, DC, pp. 245–264.

Singh, J., Goolsby, J.R., Rhoads, G.K., 1994. Behavioral and psychological conse-quences of boundary spanning burnout for customer service representatives.Journal of Marketing Research 31 (November), 558–569.

Singh, J., Verbeke, W., Rhoads, G.K., 1996. Do organizational practicesmatter in role stress processes? A study of direct and moderating effectsfor marketing-oriented boundary spanners. Journal of Marketing 60 (July),69–86.

Spector, P.E., Zapf, D., Chen, P.Y., Frese, M., 2000. Why negative affectivity should notbe controlled in job stress research: don’t throw out the baby with the bathwater. Journal of Organizational Behavior 21 (1), 79–95.

Spielberger, C.D., Vagg, P.R., Wasala, C.F., 2003. Occupational stress: job pressures andlack of support. In: Quick, J.C., Tetrick, L.E. (Eds.), Handbook of OccupationalHealth Psychology. American Psychological Association, Washington, DC, pp.185–200.

Sumer, H.C., Sumer, N., Ciftci, S., Demirutku, K., 2000. Measurement of officerpersonality attributes: a construct validity study. Turkish Journal of Psychology15 (45), 15–36 (published in Turkish).

Svensson, G., Tronvoll, B., Slatten, T., 2008. An assessment of the empirical char-acteristics of top journals in services marketing. Managing Service Quality 18(3), 289–304.

Thoresen, C.J., Kaplan, S.A., Barsky, A.P., Warren, C.R., De Chermont, K., 2003. Theaffective underpinnings of job perceptions and attitudes: a meta-analyticreview and integration. Psychological Bulletin 129 (6), 914–945.

Xanthopoulou, D., Bakker, A.B., Demerouti, E., Schaufeli, W.B., 2007. The role ofpersonal resources in the job demands-resources model. International Journalof Stress Management 14 (2), 121–141.

Van De Vliert, E., Van Yperen, N.W., 1996. Why cross-national differences in roleoverload: don’t overlook ambient temperature. Academy of Management Jour-nal 39 (4), 986–1004.

Van Dijk, P.A., Kirk, A., 2006. Emotional labor and negative job outcomes: anevaluation of the mediating role of emotional dissonance. Journal of Manage-ment and Organization 12 (2), 101–115.

Van Dyne, L., Jehn, K.A., Cummings, A., 2002. Differential effects of strain on twoforms of work performance: individual employee sales and creativity. Journal ofOrganizational Behavior 23 (1), 57–74.

Watson, D., Clark, L.A., 1984. Negative affectivity: the disposition to experienceaversive emotional states. Psychological Bulletin 96 (3), 465–490.

Watson, D., Clark, L.A., Tellegen, A., 1988. Development and validation of briefmeasures of positive and negative affect: the PANAS scales. Journal of Person-ality and Social Psychology 54 (6), 1063–1070.

Yagil, D., Luria, G., Gal, I., 2008. Stressors and resources in customer service roles:exploring the relationship between core self-evaluations and burnout. Inter-national Journal of Service Industry Management 19 (5), 575–595.

Yavas, U., Babakus, E., Karatepe, O.M., 2008. Attitudinal and behavioral conse-quences of work-family conflict and family-work conflict: doesgender matter? International Journal of Service Industry Management 19(1), 7–31.

Zapf, D., Vogt, C., Seifert, C., Mertini, H., Isic, A., 1999. Emotion work as a source ofstress: the concept and development of an instrument. European Journal ofWork and Organizational Psychology 8 (3), 371–400.