do they see eye to eye? management and employee ... · do they see eye to eye? management and...

Do They See Eye to Eye? Management and Employee Perspectives ofHigh-Performance Work Systems and Influence Processes on

Service Quality

Hui LiaoUniversity of Maryland, College Park

Keiko ToyaDoshisha University and Marketing Excellence

David P. Lepak and Ying HongRutgers, The State University of New Jersey

Extant research on high-performance work systems (HPWSs) has primarily examined the effects ofHPWSs on establishment or firm-level performance from a management perspective in manufacturingsettings. The current study extends this literature by differentiating management and employee perspec-tives of HPWSs and examining how the two perspectives relate to employee individual performance inthe service context. Data collected in three phases from multiple sources involving 292 managers, 830employees, and 1,772 customers of 91 bank branches revealed significant differences between manage-ment and employee perspectives of HPWSs. There were also significant differences in employeeperspectives of HPWSs among employees of different employment statuses and among employees of thesame status. Further, employee perspective of HPWSs was positively related to individual general serviceperformance through the mediation of employee human capital and perceived organizational support andwas positively related to individual knowledge-intensive service performance through the mediation ofemployee human capital and psychological empowerment. At the same time, management perspective ofHPWSs was related to employee human capital and both types of service performance. Finally, abranch’s overall knowledge-intensive service performance was positively associated with customeroverall satisfaction with the branch’s service.

Keywords: strategic human resource management, high-performance work systems for service quality,human capital and motivation, employee performance, customer satisfaction

A large body of strategic human resource management (HRM)research suggests that the use of high-performance work systems(HPWSs), or systems of human resource (HR) practices designed toenhance employees’ competencies, motivation, and performance, isassociated with lower employee turnover rates (e.g., Huselid, 1995),higher labor productivity (Datta, Guthrie, & Wright, 2005), lower

injury rates and better safety performance (Zacharatos, Barling, &Iverson, 2005), and better company performance (e.g., Huselid,1995). At the same time, however, several authors have raised con-cerns about some aspects of strategic HRM research. Prior researchhas primarily focused on managerial reports of the use of HPWS,ignoring the role of individual employees’ actual experiences withthese systems (Lepak, Liao, Chung, & Harden, 2006). Further, extantstrategic HRM research has predominantly taken a macro-level ap-proach and focused on establishment or firm-level outcomes; to date,there is a “dearth of research aimed at understanding how multiple (orsystems of) HR practices impact individuals [italics added]” (Wright& Boswell, 2002, p. 262). Moreover, the majority of the strategicHRM research has been conducted in manufacturing environments,neglecting the considerable presence of service factors, which nowaccount for 60% of world gross domestic product (GDP) and domi-nate economies in most nations (e.g., 71% of the GDP in Canada,73% in the United Kingdom, 74% in Japan, and 78% in the UnitedStates; The World Factbook, 2007).

The primary objective of this study was to address these issues.First, given that there may be a disconnection between whatmanagers and companies say they do as formal practices of theHPWS and what individual employees actually experience, in thisstudy we examined employee perspective with the HPWS inaddition to the management perspective. Second, we integratedmacro- and micro-level HRM research to examine the influence of

Hui Liao, Management and Organization Department, Robert H. SmithSchool of Business, University of Maryland, College Park; Keiko Toya,Graduate School of Business Sciences, Doshisha University, and Market-ing Excellence, Tokyo, Japan; David P. Lepak and Ying Hong, HumanResource Management Department, Rutgers, The State University of NewJersey.

This research was supported in part by a research grant to Hui Liao fromthe Rutgers University School of Management and Labor Relations. Anearlier version of this article was presented at the Academy of ManagementMeeting, Philadelphia, August 2007.

We thank Fumeo Maeda and Yasuhiro Kurita for their invaluable help.We thank seminar participants at the Hong Kong University of Science andTechnology; Ohio State University; Rutgers, The State University of NewJersey; University of Kansas; and University of Maryland, College Park forhelpful comments and suggestions.

Correspondence concerning this article should be addressed to Hui Liao,Robert H. Smith School of Business, University of Maryland, 4506 VanMunching Hall, College Park, MD 20742. E-mail: [email protected]

Journal of Applied Psychology © 2009 American Psychological Association2009, Vol. 94, No. 2, 371–391 0021-9010/09/$12.00 DOI: 10.1037/a0013504

371

HPWS on individual performance and to understand the psycho-logical processes through which the influence materializes. Third,we examined whether unit-level employee overall service perfor-mance translates into an important performance metric for serviceorganizations, specifically, customer satisfaction.

In what follows, we first integrate strategic HRM and the servicemanagement literature to discuss what an HPWS entails in theservice context and then discuss the importance of understandingthe system from the employees’ perspective as well as the psy-chological processes through which such a system operates toinfluence individual employees’ service performance and, ulti-mately, customer satisfaction. Figure 1 outlines the theoreticalframework of the influence of a service-quality–oriented HPWSon employee service performance.

High-Performance Work System for Service Quality

Bowen and Ostroff (2004) noted that the content of worksystems “should be largely driven by the strategic goals and valuesof the organization” and that “the foci of the human resourcemanagement practices must be designed around a particular stra-tegic focus, such as service or innovation” (p. 206). The crux ofthis statement is that to be effective, work systems must reflecthow employees add value, and this is achieved by linking thepractices within a system toward some strategic anchor. Withoutan objective, work systems lack a clear direction for employees.Therefore, all components of the HPWS should be chosen anddesigned to achieve a specific organizational objective. This stra-tegically focused approach is consistent with the argument that tobe effective, work systems should achieve horizontal fit amongvarious HR practices, such that these practices complement and arealigned with each other and achieve vertical fit, such that the work

system is aligned with the organization’s strategy (Becker &Gerhart, 1996; Huselid, 1995; Wright & McMahan, 1992).

As noted by management theorists, there are two basic strategiesfor service organizations. The first is to focus on minimizing costsand use a mass production approach in accordance with scientificTaylorism (Porter, 1980). Although reliance on technology mayhelp reduce cost and improve efficiency of service (Levitt, 1972),it may not be a sustainable advantage, as it is easily imitable and,more importantly, the low-cost pressure may create a vicious circlewhere employees and customers are increasingly dissatisfied(Schlesinger & Heskitt, 1991). The other strategy focuses onproviding high-quality service in order to enhance customer satis-faction and build a long-term relationship with customers (Gutek,1995; Porter, 1980). Happy, long-term customers “buy more, takeless of a company’s time, are less sensitive to price, and bring innew customers” (Reichheld, 1996, p. 57). Keltner (1995), forexample, found that a service-quality–focused strategy contributedsignificantly to the fact that German banks performed better thanU.S. banks in the 1980s. Other research has also shown a servicedifferentiation strategy to be associated with higher performanceof service firms (e.g., O’Farrell, Hitchens, & Moffat, 1993). Sim-ilarly, Rust, Moorman, and Dickson (2002) compared the financialreturns for three types of strategies, including (a) cost cutting, (b)revenue expansion through customer-oriented quality improve-ments, and (c) dual emphasis on cost cutting and revenue expan-sion and found that firms adopting the revenue-expansion strategyperformed the best in terms of profitability and stock returns.

The nature of services, including simultaneity of service pro-duction and consumption, intangibility of service processes andoutcomes, and customer involvement in service production(Bowen & Schneider, 1988), renders it impossible to do a quality

T1M:Management-HPWS

T2E:Employee-HPWS

T3S: Employee Human Capital

T3S: Employee Individual

Service Performance

Individual Level

Employment Group Level

T3E: Employee Psychological Empowerment

T3E: Employee Perceived

Organizational Support

Figure 1. Theoretical framework of the influence processes of the high-performance work system (HPWS) forservice quality on employee service performance and data collection schedule and sources. The horizontaldashed line separates employment group-level and individual-level constructs; the arrow crossing a dashed linefrom upper level to lower level represents cross-level, top-down effect on the lower level outcome variables.Management-HPWS: Management’s perspective of the HPWS practices generally in use for an employee groupof certain employment status. Employee-HPWS: Employee individually experienced HPWS. T1M � Time 1management survey; T2E � Time 2 employee survey; T3E � Time 3 employee survey; T3S � Time 3supervisor evaluations.

372 LIAO, TOYA, LEPAK, AND HONG

control check after production to ensure quality as in the manu-facturing setting (Schneider, White, & Paul, 1998). Therefore, theperformance of front-line employees, or their behaviors of helpingand serving customers to address customer needs (Liao & Chuang,2004), directly influences customer satisfaction with the servicequality. In order for front-line employees to provide high-qualityservice, organizations need to design a work system that ensuresthat employees have the knowledge, skills, and abilities, as well asthe motivation, to meet customer needs. A few studies haveexplicitly examined the linkage between HRM practices and ser-vice quality. Schneider et al. (1998) proposed that service qualityrests on a set of organizational “foundation issues” that supportand facilitate front-line employee service delivery, which includeinternal service provided by support staff, efforts to remove ob-stacles to work, and employee participation and training. Usingdata from bank branches, Schneider et al. found that the foundationissues were positively associated with branch service climate,which was positively associated with customer evaluations ofservice quality. Batt (2002) found that high-involvement practicescharacterized by high skills, discretion, and incentives were asso-ciated with lower quit rates and subsequently higher sales growthof call centers. Liao and Chuang (2004) examined three HRMpractices and found that employee involvement in decision makingand service training were positively related to restaurant employ-ees’ service performance, which in turn was positively related tocustomer satisfaction and loyalty.

Building on these studies as well as the frameworks of HPWSsdescribed by Pfeffer (1998) and Zacharatos et al. (2005), wepropose an HPWS for service quality and define it as a system ofHR practices designed to enhance employees’ competencies, mo-tivation, and performance in providing high-quality service toexternal customers. In this view, HPWS includes practices ofextensive service training, information sharing, self-managementservice teams and participation, compensation contingent on ser-vice quality, job design for quality work, service-quality–basedperformance appraisal, internal service, service discretion, selec-tive hiring, employment security, and reduced status differentia-tion.1 This conceptualization of the HPWS for service qualitycaptures the foundation HRM issues deemed important for servicedelivery in Schneider et al.’s (1998) framework and includes theHR practice dimensions examined in prior strategic HRM studiesin the service settings (Batt, 2002; Delery & Doty, 1996). In thecurrent study, these practices are anchored with a goal to promoteservice quality. For example, extensive training emphasizes edu-cating employees on how to provide quality service; performanceappraisal uses service criteria; and contingent compensation linkspay to service quality. These work practices, taken together, pro-vide employees with the knowledge, skills, and abilities; re-sources; information; and discretion they need to meet customerdemands, as well as the motivation to provide high-quality service.

Understanding the Employee Perspectiveof the High-Performance Work System

Focusing on employees as providers of services also dictates thatparticular attention be paid to employees’ actual experience of theHPWS for service quality. Wright and Boswell (2002) noted that“much of the macro HRM research empirically assumes invari-ability in HR practices across large groups of jobs within organi-

zations” (p. 264) and has focused primarily on management per-spective of the HR practices generally implemented for all of theemployees in an organization. Few studies have examined theHPWS targeted to different employment groups, and even fewerhave examined the HPWS actually experienced by individualemployees. We argue that this is an oversight for several reasons.

First, different employee groups may not have identical expe-riences of HR practices. There is a long-standing tradition ofvariance in exposure to HR practices, such as compensation,training, and promotion opportunities across administrative bound-aries such as exempt versus nonexempt status or managerial versusnonmanagerial status (Huselid, 1995). Companies may also usedifferent HR practices to match the requirements of particularemployee groups (Miles & Snow, 1984; Osterman, 1987). Forexample, Lepak and Snell (2002) showed that core employeesreceived significantly greater exposure to a commitment-orientedwork system and that noncore employee groups tended to bemanaged by work systems that convey lower levels of investmentin employees. Relatedly, Melian-Gonzalez and Verano-Tacoronte(2006) found that the work systems used for core employees weremore sophisticated than those used for other employee groups.Similarly, Lepak, Taylor, Tekleab, Marrone, and Cohen (2007)demonstrated that establishments provided higher levels of expo-sure to high-investment work systems for core employees, ascompared with support employees, and that the differences weremagnified in nonmanufacturing environments. The logic for differ-entiating work systems across employee groups is that work systemsare used to match the level of investment needed to reflect therelative status of different employees groups and maximize theirpotential contributions to competitive success. Building on thislogic and existing research, we anticipate that employees in groupsof different employment status will experience differences in theirexposure to the HPWS practices. Such between-group difference isexpected to contribute to the variability in employees’ experienceswith the HPWS.

Second, even within the same employee group, members whotheoretically should share the same HPWS practices may betreated differently or have different perceptions or experiences ofthe practices in place. This within-group difference is anothersource of the variability in employees’ experiences with theHPWS. For instance, the organizational diversity literature hasrevealed earning differences between women and men and be-tween White persons and people of color who are similarly qual-ified and working in the same job within the same organization,suggesting the existence of differential treatment toward differentemployees in management practices (e.g., Joshi, Liao, & Jackson,2006; Pfeffer & Davis-Blake, 1987). The organizational justiceliterature has suggested that employees often form different eval-uations on the allocations of resources, such as pay and promotion,the procedures used to arrive at these allocations, and the inter-personal treatment and information given to the employees asthese procedures are carried out (e.g., see Colquitt, Conlon,Wesson, Porter, & Yee, 2001, for a review). Further, the leader–

1 We excluded the practices of selective hiring, employment security,and reduced status differentiation in our empirical testing of this study dueto the nature of the sample. See the Measures section for a detailedexplanation.

373HIGH-PERFORMANCE WORK SYSTEMS

member exchange (LMX; Graen & Uhl-Bien, 1995) literature hasproposed that leaders establish different social exchange relation-ships with different subordinates; employees who have a high-quality LMX with their supervisor have the advantages of ampleresources, more training opportunities, premier assignments, emo-tional support, decision-making responsibilities, and cooperativeinteractions with the supervisor (Liden & Graen, 1980). Consistentwith this view, Rousseau (2005) noted that individual employeesnegotiate their idiosyncratic employment arrangement, or“I-deals,” with employers. In addition, Bowen and Ostroff (2004)noted that

HRM practices can be viewed as a symbolic or signaling function bysending messages that employees use to make sense of and to definethe psychological meaning of their work situation (e.g., Rousseau,1995). All HRM practices communicate messages constantly and inunintended ways, and messages can be understood idiosyncratically,whereby two employees interpret the same practices differently(Guzzo & Noonan, 1994). (p. 206)

Taken together, these theoretical perspectives and empiricalfindings imply that HR practices may be implemented differentlyfor employees, or at least that employees may perceive or experi-ence differences in exposure to work practices. The lack of uni-formity across employees in their experiential-based perceptionsabout the HPWS practices suggests that there will exist a discon-nection between what management says about the HPWS practicesgenerally implemented for a particular employee group (hereinaf-ter referred to as management-HPWS) and the HPWS practicesactually experienced by the individual employees in that group(hereinafter referred to as employee-HPWS). Therefore, it is nec-essary to examine employee-HPWS in addition to management-HPWS to understand the psychological process through which formalpractices of HPWS influence individual employees’ human capital,motivation, and behaviors in service delivery.

Influence of Employee Perspective of theHigh-Performance Work System on Individual

Service Performance

Next, we examined the influence process of employee-HPWSon employee performance. We argue that the employee-HPWS ispositively related to employee service performance by enhancingemployee human capital and motivation required for service de-livery. Figure 1 depicts the proposed mediation relationships. Ourfocus on these mediating factors is in line with the prominenttheory of job performance (Campbell, McCloy, Oppler, & Sager,1993), which suggests that the direct determinants of individualperformance are individuals’ human capital and motivation andthat other individual differences, such as experiences and person-ality, and contextual factors such as HR interventions, indirectlyinfluence job performance through their impact on the individuals’human capital and motivation.

Human Capital

Human capital refers to employee knowledge, skills, and abili-ties that are valuable for the firm (e.g., Subramaniam & Youndt,2005). Strategic HRM research has used human capital as a the-oretical underpinning (Jackson & Schuler, 1995), with the argu-

ment that one important function of HRM is its “buying” and“making” of desirable employee knowledge, skills, and abilities,which can in turn be used to create value for the firm (e.g., Becker& Huselid, 1998; Delery & Shaw, 2001; Lado & Wilson, 1994;Snell, Youndt, & Wright, 1996). Despite its wide mention inempirical strategic HRM studies (e.g., Batt, 2002; Huselid, 1995),the role of human capital has rarely been examined explicitly as amediator between work systems and performance outcomes.

In a service context, employees need to have a good knowledgeabout the services, products, and customer needs and to have theabilities and skills to meet customer needs. We argue thatemployee-HPWS, a key mission of which is to select and developservice talents through practices such as service-quality–focusedhiring, training, information sharing, performance feedback, andso forth may enhance employee human capital for service deliveryand subsequently service performance. Therefore, we propose thefollowing:

Hypothesis 1: The positive relationship between employee-HPWS and employee service performance is mediated by em-ployee human capital.

Motivation

Motivation refers to an individual’s direction, intensity, andduration of effort (Campbell et al., 1993). Whereas human capitalprovides the capabilities for employees to contribute, motivationdeals with the extent to which employees are willing to utilizethese capabilities. Jackson and Schuler (1995) pointed out that “thepotential value of human capital can be fully realized only with thecooperation of the person” (p. 241). HRM practices need to effec-tively align the interests of employees and employers so thatemployees are willing to exert their effort (Delery & Doty, 1996).Yet, motivation has seldom been measured explicitly or tested instrategic HRM studies. In contrast, the importance of motivationhas long captured psychology researchers’ attention. Two motiva-tion constructs that are particularly related to individual employeejob performance are psychological empowerment and perceivedorganizational support.

Psychological empowerment. Psychological empowerment re-fers to individuals’ self-motivating mechanisms and consists ofmeaning, competency, self-determination, and impact (Spreitzer,1995; Thomas & Velthouse, 1990). Meaning is an individual’sperceived value of work compared with the individual’s personalgoals and standards; competence is an individual’s confidence inhis or her ability to perform the work satisfactorily; self-determination is an individual’s sense of control in initiatingand changing actions; and impact is an individual’s perceivedinfluence over important strategic, administrative, or operatingoutcomes.

Although psychological empowerment reflects individuals’ in-nate intrinsic task motivation, it can be influenced by externalpractices (Seibert, Silver, & Randolph, 2004; Spreitzer, 1995). Weargue that HPWS may represent such empowering work practices.For examples, service-quality–focused performance feedback andinformation sharing may help employees perceive the service tasksto be meaningful and important; extensive service training anddelegation of decision-making power may enhance employees’confidence in their competence in service delivery; and increased


discretion in customizing service delivery and handling customercomplaints may increase employees’ perceived self-determinationand impact.

We further argue that psychological empowerment will lead tobetter service performance. Research shows that psychologicalempowerment leads to employee commitment to and internaliza-tion of task goals (Kanter, 1983), persistence of effort in nonrou-tine situations and resilience in adversary situations (Bandura,1977), initiative (Thomas & Velthouse, 1990), and innovativebehaviors (Spreitzer, 1995). All of these characteristics may facili-tate service performance, which requires employees to learn about thediverse needs of customers, handle the uncertainty customers intro-duce when participating in service, and adapt their interpersonalstyle and service offering to customer needs. Indeed, Gwinner,Bitner, Brown, and Kumar (2005) found intrinsic motivation to bepositively related to employee adaptive service behaviors.

Whereas prior studies have demonstrated the positive impact ofintrinsic motivation on performance criteria, the evidence concern-ing the role of extrinsic motivation is less convincing. Cognitiveevaluation theory (Deci, 1975) suggests that extrinsic motivationregulates a behavior by way of external contingencies, diminishesfeelings of autonomy, and activates a change in perceived locus ofcausality from internal to external, hence undermining intrinsicmotivation. Ryan, Mims, and Koestner (1983), on the contrary,found that extrinsic rewards contingent on high-quality perfor-mance given in a supportive interpersonal context actually en-hanced intrinsic motivation. However, Gwinner et al. (2005) foundextrinsic motivation to be unrelated to employee customized ser-vice delivery. Given these inconclusive arguments and findings,we focused on intrinsic motivation in our hypothesis developmentand controlled for extrinsic motivation in our analyses.

Hypothesis 2: The positive relationship between employee-HPWS and employee service performance is mediated byemployee psychological empowerment.

Perceived organizational support. Beyond psychological em-powerment, we propose that another reason for employees to bemotivated by the HPWS is a favorable social exchange with theorganization. On the basis of the social exchange theory (Blau,1964), Settoon, Bennett, and Liden (1996) argued that “positive,beneficial actions directed at employees by the organization and/orits representatives contribute to the establishment of high-qualityexchange relationships that create obligations for employees toreciprocate in positive, beneficial ways” (p. 219). Perceived orga-nizational support, or employees’ perceptions of the extent towhich organizations value employees and care about their well-being (Eisenberger, Fasolo, & Davis-Mastro, 1990), is essential informing such obligations (Shore & Wayne, 1993). Researchershave argued that when employees perceive a high level of orga-nizational support, they may use behaviors valued by the organi-zation as currency to reciprocate the benevolent treatment from theorganization (Lambert, 2000; Shore et al., 2004). Similarly,Schneider and Bowen (1985) argued that service employees arelikely to treat customers well if the organization treats the employ-ees well.

Numerous efforts have examined the antecedents and outcomesof perceived organizational support. Such support has been shownto be influenced by organizations’ investment in employees

through HR practices, such as training and development, andorganizations’ recognition of individual achievement throughpractices such as promotions and salary increases (e.g., Wayne,Shore, Bommer, & Tetrick, 2002; Wayne, Shore, & Liden, 1997).In addition, perceived organizational support is found to play a keyrole in relating to employees’ commitment to the organization(Eisenberger et al., 1990; Settoon et al., 1996; Wayne et al., 1997),satisfaction (Shore & Tetrick, 1991), conscientiousness in per-forming tasks and innovation (Eisenberger et al., 1990), loyalty(Tetrick, Shore, Newton, & Vandenberg, 2007), and organizationalcitizenship behavior and lower intention to quit (Wayne et al.,1997).

Despite these efforts, to our knowledge, few strategic HRMstudies have integrated perceived organizational support as a link-age between HPWS and employee performance. One exceptionwas the study by Allen, Shore, and Griffeth (2003), which showedthat service employees’ perceptions of participation in decisionmaking, fairness of rewards, and growth opportunities were posi-tively associated with their development of perceived organiza-tional support, which, in turn, was positively associated with theirjob satisfaction and organizational commitment, as well as lowerturnover.

Building on these rationales and findings, we argue that per-ceived organizational support may be a potential path by whichHPWS influences employee performance in a service context.Practices such as extensive training reflect companies’ investmentin employees; practices such as internal career opportunities, ser-vice discretion, and self-management service teams demonstratemanagement’s respect of employees’ opinions and initiatives; andpractices such as service-quality–based performance appraisal andcompensation show the recognition of employee service excel-lence. Therefore, we propose the following:

Hypothesis 3: The positive relationship between employee-HPWS and employee service performance is mediated byemployee perceived organizational support.

Influence of Management Perspectiveof High-Performance Work Systems on

Individual Outcomes

Thus far we have proposed hypotheses regarding the influencesof employee-HPWS on individual outcomes. As noted above,employee-HPWS may differ from management-HPWS. However,this is not to suggest that there is no relationship between these twoperspectives. We argue that as management-HPWS represents theHPWS practices generally implemented for a particular group ofemployees, to a certain extent it reflects the objective environmentshaped by formal management practices. As a result, management-HPWS provides a contextual cue for employees to form theirperceptions and experience of the work system. Therefore, weexpect management and employee perspectives of the HPWS to bepositively related.

At the same time, however, employee-HPWS may have a moreproximal relationship with employee individual outcomes, because itis the employees’ actual experiences and perception of the context,not the context itself nor the cues obtained from the context, thatdirectly determine their reactions (James, James, & Ashe, 1990;Schneider, 1990). Therefore, management-HPWS affects an em-


ployee to the extent that it affects the employee’s experience andperception of the HPWS. As a result, we propose that employee-HPWS mediates the effects of management-HPWS on employeehuman capital and motivation.

Hypothesis 4: The positive relationship between management-HPWS and individual employee human capital, psychologicalempowerment, and perceived organizational support is medi-ated by employee-HPWS.

Taken together, Hypotheses 1 through 4 suggest thatmanagement-HPWS influences employee-HPWS, which, in turn,impacts individual employees’ human capital, psychological em-powerment, and perceived organizational support, and these fac-tors further influence individual employee service performance.

Linking Employee Service Performanceto Customer Satisfaction

Next, we propose that employee service performance may fur-ther influence customer satisfaction on two levels. First, at theindividual level, employee service performance to a particularcustomer, or the one-on-one interaction between an employee anda customer, may directly affect the customer’s satisfaction withthis particular service encounter and the customer’s decision re-garding whether to continue to use the service of this particularemployee (Gutek, 1995; Gutek, Bhappu, Liao-Troth, & Cherry,1999). For example, by observing 191 bank tellers, Pugh (2001)found that individual tellers’ display of positive emotions wasdirectly associated with their customers’ positive affect and sub-sequently with positive evaluations of service quality. Liao andChuang (2007) found that a hairdresser’s individual-level serviceperformance directly determines customers’ willingness to have along-term service relationship with the hairdresser and the numberof long-term customers the hairdresser maintains. Therefore, anindividual employee’s service performance may directly affectcustomer satisfaction.

Second, a customer’s overall experience with a service unit maybe shaped by the customer’s multiple encounters with multipleservice employees in the unit (Liao & Chuang, 2004). For exam-ple, a banking customer may likely interact with different front-line employees on different visits and for different banking orinvestment needs. Thus, at the business-unit level of analysis, theoverall service performance across the front-line employees mayaffect customer satisfaction with the unit’s overall service delivery.Whereas theoretically both individual-level and unit-level em-ployee service performance may affect customer satisfaction, test-ing the individual-level effect requires an exact match between theemployee and the customers he or she serves (see Liao & Chuang,2007, for such a design and test). The current study did not includesuch matched data but instead assessed the individual customers’satisfaction with the unit’s overall service. Therefore, with fullacknowledgment of the theoretical and business relevance for theeffect of individual-level employee service performance, in thecurrent study we focused on the effect of unit-level employeeservice performance and proposed the following:

Hypothesis 5: A unit’s overall employee service performanceis positively related to customer satisfaction with the unit’soverall service quality.

Method

Study Design and Participants

We tested the proposed theoretical framework using data from anational bank in Japan. All 92 branches of the bank participated inthe study. In order to link information from multiple stakeholdersinvolved in the service profit chain, test the temporal linkagesamong the study variables, and reduce common method bias, wecollected information from five sources (headquarters, branch se-nior managers, employee supervisors, customer-contact employ-ees, and branch customers), in two formats (surveys and archivaldata), and in three phases.

Specifically, at Time 1, branch management (including branchsenior managers and employee supervisors) filled out a surveyabout the HPWS practices implemented in their branch for threeemployee groups of different employment status (group differ-ences are explained in the Measures section). At Time 2, individ-ual employees filled out a survey about their personal experiencesof the work system. At Time 3, employees filled out anothersurvey on variables including their psychological empowerment,extrinsic motivation, and perceived organizational support. Weseparated employee measures into these two surveys at differenttimes to reduce common-method bias. In addition, at Time 3,employee supervisors provided evaluations of each individual em-ployee’s human capital and service performance. The time intervalbetween two adjacent phases ranged from 2 to 4 weeks. Concur-rent with the data collection within the branches, customer satis-faction measures were collected from the branch’s external cus-tomers, and employee demographic information and branchcharacteristics, such as size, age, and local competition, wereobtained from the bank headquarters’ archival data.

To solicit honest responses, we arranged for the marketingconsulting company run by Keiko Toya to administer the entiredata collection. All branch senior managers, supervisors, and em-ployees of the bank were invited to participate in the study, and astratified random sample of each branch’s external customers wasselected to represent customers of different asset levels. The mar-keting consulting company precoded the surveys with branch andindividual identification numbers and sent all of the surveys di-rectly to the participants’ home together with a postage-paid returnenvelope addressed directly to Keiko Toya. The respondents wereassured of confidentiality and that nobody from the bank wouldhave access to their individual responses. To further reduce po-tential psychological stress, we did not include any question in theemployee surveys pertaining to individuals’ names, locations, orpersonal demographic information and obtained these data directlyfrom the bank’s headquarters. Upon returning their completedsurvey, the customers each received a payment in Japanese yenequivalent to about 10 U.S. dollars, while the bank’s managers,supervisors, and employees did not receive any payment.

With endorsement from the bank’s top management, weachieved high response rates for the branch manager and employeesamples across the branches. A total of 292 (91%) branch manag-ers filled out the Time 1 management survey, 1,149 (92%) em-ployees filled out both of the two employee surveys, and 223supervisors provided performance evaluations for 959 (77%) ofthe employees. We had a final usable sample with completematched management–supervisor–employee information for 830employees from 91 branches. We compared the 830 employees


included in the final sample with the excluded employees havingincomplete information and found that the final sample had higherfemale representation and were younger than the unused sample.To account for the potential effects of gender and age, we con-trolled for these variables in all analyses. In addition, we obtained1,772 usable customer responses from 75 branches, representingan overall response rate of 27%, which is comparable to the surveyresponse rate from random survey sampling reported in the mar-keting literature (e.g., Seiders, Voss, Grewal, & Godfrey, 2005).We conducted t tests of two-sample means and found that, incomparison with these 75 branches, the 16 branches with nocustomer data had fewer female employees and more competingbanks in the neighborhood but no statistically significant differ-ence for any of the other study variables. We controlled for thesevariables in all analyses.

Survey Translation Procedures

For measures that were originally in English, we followed aniterative translation procedure. First, Hui Liao worked closely witha Japanese linguist who teaches Japanese at a U.S. university totranslate the surveys from English to Japanese. Second, KeikoToya, who is a Japanese native and proficient in English, teachesmarketing at a university in Japan, and has extensive consultingexperience with the Japanese banking industry, checked the trans-lation for accuracy, discussed the relevance and appropriateness ofthe questions in depth with focus groups of managers and employ-ees from the bank, identified areas of concerns, and suggestedquestions to be deleted, added, and modified. Third, the Japaneselinguist, in consultation with Hui Liao and Keiko Toya, worked toresolve these issues. Fourth, Keiko Toya’s business partner, who isa Japanese native with a U.S. master’s in business administrationdegree and consults with the Japanese banking industry, improvedthe readability of the questions through discussions with KeikoToya.

Next, we describe the measures for the main study variablesinvolved in the analyses of employee service performance andcustomer satisfaction, respectively, by each level of analysis. Wethen describe the measures for the control variables by each levelof analysis.

Measures of Individual-Level Variables Relevant forPredicting Employee Service Performance

Employee-experienced high-performance work system. Theemployee measure assessed employees’ individual experiences ofthe work system. Employees answered questions on the basis oftheir personal experience and understanding of the HR practices ona 5-point Likert scale ranging from 1 (strongly disagree) to 5(strongly agree). Our measure of HPWS included eight practicedimensions that aim to enhance employee human capital, psycho-logical empowerment, and obligation in delivering high-qualityservice. Table 1 reports the practice dimensions, scale sources,means, standard deviations, and coefficient alphas.

The first six practice dimensions were based on Zacharatos etal.’s (2005) framework, and the measures were primarily derivedfrom the employee version of the HPWS used in Study 2 ofZacharatos et al. and adapted in the current study to have aservice-quality focus. The dimensions include extensive service T

able

1H

igh-

Per

form

ance

Wor

kSy

stem

s(H

PW

Ss):

Subs

cale

s,E

mpl

oyee

Per

spec

tive

(Em

ploy

ee-H

PW

S),

and

Man

agem

ent

Per

spec

tive

(Man

agem

ent-

HP

WS)

for

Dif

fere

ntE

mpl

oyee

Gro

ups

Prac

tice

dim

ensi

on

Em

ploy

ee-H

PWS

(��

.82)

Man

agem

ent-

HPW

Sfo

rem

ploy

eegr

oup

1(�

�.8

6)

Man

agem

ent-

HPW

Sfo

rem

ploy

eegr

oup

2(�

�.8

6)

Man

agem

ent-

HPW

Sfo

rem

ploy

eegr

oup

3(�

�.8

4)

Scal

eso

urce

MSD

�L

oadi

ngM

SD�

Loa

ding

MSD

�L

oadi

ngM

SD�

Loa

ding

Ext

ensi

vese

rvic

etr

aini

ng3.

130.

61.7

6.5

13.

520.

58.8

0.5

43.

510.

55.8

1.5

22.

970.

71.8

4.4

6D

eler

y&

Dot

y(1

996)

;Z

acha

roto

set

al.

(200

5)In

form

atio

nsh

arin

g3.

530.

63.8

4.7

93.

960.

50.8

3.6

83.

850.

51.8

2.6

33.

120.

68.8

5.6

5Z

acha

roto

set

al.

(200

5)Se

lf-m

anag

edte

ams

and

part

icip

atio

n3.

370.

61.8

0.5

53.

700.

55.7

4.7

03.

680.

50.7

2.7

33.

310.

61.7

7.8

0Z

acha

roto

set

al.

(200

5)C

ompe

nsat

ion

cont

inge

nton

serv

ice

perf

orm

ance

2.96

0.61

.78

.81

3.19

0.64

.64

.51

3.13

0.61

.66

.41

2.73

0.61

.78

.47

Zac

haro

tos

etal

.(2

005)

Job

desi

gn3.

630.

58.7

2.5

13.

770.

52.5

1.5

83.

790.

51.5

8.6

52.

950.

69.6

7.6

9Z

acha

roto

set

al.

(200

5)Se

rvic

e-qu

ality

-bas

edpe

rfor

man

ceap

prai

sal

3.09

0.68

.87

.52

3.82

0.70

.87

.52

3.85

0.58

.81

.51

3.61

0.70

.84

.41

Del

ery

&D

oty

(199

6);

Zac

haro

tos

etal

.(2

005)

Inte

rnal

serv

ice

3.46

0.77

.73

.56

3.82

0.59

.81

.52

3.80

0.57

.82

.52

3.66

0.61

.83

.58

Schn

eide

ret

al.

(199

8)Se

rvic

edi

scre

tion

2.83

0.67

.86

.59

3.15

0.63

.76

.59

2.95

0.66

.81

.66

2.72

0.72

.85

.73

Dev

elop

edfo

rpr

esen

tst

udy


training (6 items, e.g., “The branch supports me to join the cus-tomer service training program provided by the headquarters”),information sharing (8 items, e.g., “Customers’ suggestions onhow to improve service quality is shared with me”), self-management service teams and participation (6 items, e.g., “Sug-gestions for improving customer service from employees like meare usually implemented in full or in part within this branch”),compensation contingent on service quality (8 items, e.g., “My payis tied to the quality of service I deliver to the customers”), jobdesign for quality work (5 items, e.g., “My job is designed to besimple and repetitive”; reverse-coded), and service-quality–basedperformance appraisal (4 items, e.g., “To what extent does yourbranch evaluate your performance based on a track record of yourcourteous service to customers”). Several items deemed irrelevantfor the Japanese banking industry were dropped from Zacharatoset al.’s (2005) original scale (e.g., “Such things as my previousinjuries and my alcohol and substance use were assessed before Iwas hired to work here”). We also supplemented Zacharatos etal.’s measures of the training and performance appraisal practiceswith items adapted from Delery and Doty’s (1996) measure ofwork systems for banking industries (e.g., “The training programsI went through in this branch prepared me to provide high-qualityservice”).

In addition to these six dimensions, we included two practicesdeemed important for the service context. Internal service refers tothe interdepartmental support provided to service employees andhas been argued to influence front-line employees’ service deliv-ery to external customers (Schneider et al., 1998). We used the twoitems reported in Schneider et al. (1998) to assess internal service(e.g., “Employees in the other departments of this branch cooper-ate well with me to get my job done”). Service discretion refers tothe level of authority employees have in resolving customer com-plaints and customizing service offering and has been argued toinfluence employee prompt and responsive handling of customercomplaints and employee adaptive service behaviors (Gwinner etal., 2005). We developed five items to assess service discretion(e.g., “I have the discretion to customize the service offering tomeet customer needs”).

We dropped selective hiring, which was included in Zacharatoset al. (2005), because the bank has centralized staffing at theheadquarters and thus the branches do not hire their own employ-ees. We also dropped employment security and reduced statusdifferentiation dimensions, included in Zacharatos et al., becauseas we explain later, as a standard practice across branches, thebank has clear hierarchical status differentiation for differentgroups of employees, and whether an employee has lifetime em-ployment security depends on the group to which the employeebelongs. Therefore, there would be no between-branches variancein formal management practices in terms of these three dimen-sions. In addition, because a goal of the current study was toexamine the same set of HR practices from both the employee andmanagement perspectives, we excluded these three dimensionsfrom the measures of the employee perspective.

We then calculated a unitary index of employee-HPWS. Thisindex approach has been recommended and widely used in stra-tegic HRM research, as it is consistent with one of the fundamentalprinciples of strategic HRM research, which argues that the impactof HR practices is best understood by examining the system of HRpractices in place instead of examining HR practices individually

(Becker & Huselid, 1998; Delery, 1998; Guthrie, 2001; Huselid,1995; Lepak et al., 2006; Ostroff & Bowen, 2000; Wright &Boswell, 2002). We followed the subscale aggregation approach,as supported by Drasgow and Kanfer (1985) and used in priorstrategic HRM studies such as Zacharatos et al. (2005). We firstcalculated the subscale scores, averaging across items of the samepractice dimension (e.g., selective hiring), which was justified bythe high internal consistency of the subscales; we then created theindex of employee-HPWS, averaging across the eight practicedimensions, which again was justified by an alpha of .89 across theHR practices. Further, we conducted a principal factor analysis forthe practices included in employee-HPWS. Only one factor had aneigenvalue of greater than 1, providing additional support for theunitary-index approach. Table 1 reports the factor loadings.

Employee psychological empowerment. In the second surveysent to employees, employees reported their individual psycholog-ical empowerment using Spreitzer’s (1995) 12-item scale adaptedto the service context. The 5-point Likert scale (1 � stronglydisagree; 5 � strongly agree) assessed employee perceived mean-ing (e.g., “The service work I do is meaningful to me”), compe-tency (e.g., “I am self-assured about my capabilities to perform myactivities in customer service”), self-determination (e.g., “I haveconsiderable opportunity for independence and am free in how I domy job in serving customers”), and impact (e.g., “I have significantinfluence over what happens in my branch”) in service delivery.Alpha was .89.

Employee perceived organizational support. In the secondemployee survey, employees evaluated their perceived organiza-tional support from the branch, using the eight-item perceivedorganizational support scale of Eisenberger, Cummings, Armeli,and Lynch (1997). Respondents indicated their agreement witheach of the statements using a 7-point scale ranging from 1(strongly disagree) to 7 (strongly agree) (e.g., “My branch caresabout my well-being”). Alpha was .89.

Employee human capital. An employee’s human capital wasevaluated by his or her direct supervisor using a five-item humancapital scale by Subramaniam and Youndt (2005) and Youndt,Subramaniam, and Snell (2004). Participants responded on a7-point scale ranging from 1 (strongly disagree) to 7 (stronglyagree). The items were adapted to describe service related knowl-edge, skills, and abilities (e.g., “This employee is highly skilled inserving customers”). Alpha was .94.

Employee service performance. Employees’ direct supervisorsevaluated two types of service performance for each employee ona 7-point scale (1 � highly unsatisfactory; 7 � highly satisfac-tory): general service performance and knowledge-intensive finan-cial service performance. General service performance refers tothe overall professional appearance and the reliability, responsive-ness, assurance, and empathy displayed by employees in servingcustomers. It was assessed by 23 items adapted from the widelyused service quality scale of Parasuraman, Zeithaml, and Berry(1994; e.g., “how satisfactory is this employee’s performanceregarding providing services as promised?”). General service per-formance is applicable for all service sectors, and these servicequality criteria have been used in numerous marketing studies (seeAsubonteng & McCleary, 1996, for a review). Knowledge-intensive service performance refers to the financial service thatrequires specialized professional knowledge and skills and is onlyrelevant for the financial industry. We developed an 18-item scale


to assess employee performance in recommending, designing, andselling appropriate investment products, flexible/fixed-returnproducts, pension plan products, loan products and mortgageproducts, managing customer tangible and intangible assets,handling irregular cases, and so forth. Alpha was .97 for generalservice performance and .96 for knowledge-intensive serviceperformance.

Because employee supervisors provided the ratings for individ-ual employees’ human capital, general service performance, andknowledge-intensive service performance in one survey and be-cause knowledge-intensive service performance is a new scale, weconducted a series of confirmatory factor analyses to examine thediscriminant validity of these three constructs. We first createditem parcels for each scale in order to yield more stable parameterestimates (Kishton & Widaman, 1994; Yuan, Bentler, & Kano,1997). For example, the 23 items of the general service perfor-mance scale were randomly assigned to five parcels, and theaverage item scores of the five parcels were used as indicators forgeneral service performance. We found that the one-factor modelfit the data poorly, �2(54) � 4,218.44, non-normed fit index(NNFI) � .77, comparative fit index ( CFI) � .82, incremental fitindex ( IFI) � .82, and root mean square residual ( RMSR) � .19.In contrast, the hypothesized three-factor model fit the data well,�2(51) � 757.28, NNFI � .96, CFI � .97, IFI � .97, RMSR �.045, and significantly better than the one-factor model, ��2(3) �3,461.15, p � .001; better than an alternative two-factor modelcombining human capital and general service performance intoone factor, ��2(2) � 315.39, p � .001; better than an alternativetwo-factor model combining human capital and knowledge-intensive service performance into one factor, ��2(2) � 3,284.2,p � .001; and better than an alternative two-factor model combin-ing general service performance and knowledge-intensive serviceperformance into one factor, ��2(2) � 3,130.62, p � .001. Fur-ther, we constrained the estimated correlation parameter (�ij)between two of the three variables to 1.0 at a time and thenperformed chi-square difference tests on the values obtained forthe constrained and unconstrained models (Anderson & Gerbing,1988). All of the three chi-square difference tests revealed that theunconstrained models fit the data significantly better than did theconstrained model, suggesting that the imposed constraint wasunrealistic. In addition, we used the test recommended by Ander-son and Gerbing (1988) as well as Bagozzi, Yi, and Phillips (1991)to examine whether the 95% confidence interval around the cor-relation of each pair of the factors contained the value of one (�1or �1). We found that none of the 95% confidence intervalscontained the value of one. Taken together, these results provideddiscriminant validity evidence that the three factors are distinctfrom one another.

Measures of Group-Level Variables Relevant forPredicting Employee Service Performance

To measure management’s assessment of the HPWS practicestargeted to different employee groups at each branch, we asked allmanagers to fill out a survey consisting of three sections; the HRpractice questions repeated in each section for three differentgroups of employees. In the bank evaluated in this study, employ-ees are categorized into three groups with different official status,employment security, and advancement opportunities. Group 1, or

“Sougoshoku” employees, is the group with the highest status;employees in this group enjoy a lifetime contract, have unlimitedtransferability to branches of their choice and promotion opportu-nities to management, and are mostly men. Members of Group 2,or “Ippanshoku” employees, have lower status, maintain a lifetimecontract but limited transferability and fewer promotion opportu-nities to management, and are mostly women. Those in Group 3,or part-time employees, have the lowest status, no lifetime con-tract, and no transfer or promotion opportunities and are predom-inantly women. All three groups of employees have direct contactwith customers. Because management may have different practicestargeted to different groups, we measured management perspec-tives of HPWS for each group separately.

We conceptualize management-HPWS as an eight-dimensionsystem corresponding to the employee-HPWS measures. Theitems were primarily adapted from the management version usedin Study 1 of Zacharatos et al. (2005) and supplemented with itemsfrom Delery and Doty (1996) and Schneider et al. (1998), exceptfor the service discretion dimension, which, as noted above, wasmostly created for this study. On a 5-point scale ranging from 1(strongly disagree) to 5 (strongly agree), branch managers an-swered questions about the formal HR practices for each of thethree employee groups.

We then calculated the index scores of management-HPWSfor the three groups of employees following a similar procedureused in calculating the employee-HPWS index score. The highinternal consistency both within each practice dimension andacross the eight dimensions, as well as the overall one-factorstructure of the practices, empirically justified the index ap-proach (see Table 1). In each branch, an average of threemanagers reported management-HPWS targeted toward the em-ployee groups.2 We aggregated the index scores across themanagers of the same branch to form the measure of manager-HPWS for each employee group. Aggregation is justified by ahigh interrater agreement (rwg(j); James, Demaree, and Wolf,1984) that adjusted for a slight negative skew in the expectedvariance; the median rwg(j) value was .97 for management-HPWS targeted toward Group 1 employees, .98 targeted towardGroup 2 employees, and .97 targeted toward Group 3 employ-ees. In addition, ICC1(intraclass correlation), or the proportionof between-branch variance in the total variance, was signifi-cant for all three types of measures, with the values of .16, .11,and .17, respectively, which are comparable to the inflatedmedian ICC1 value of .12 reported in the organizational liter-ature (see Bliese, 2000). Further, ICC2, or the reliability ofbranch mean, was .38, .28, and .34, respectively. The relativelylow ICC2 values are a direct result of the low ICC1 values andthe very small number of managers from each branch. LowICC2 values suggest that it may be difficult to detect emergentrelationships using branch means (Bliese, 2000); however, theyshould not prevent aggregation if aggregation is justified bytheory and supported by high rwg(j) (Chen & Bliese, 2002;

2 In our sample, 9 out of the 91 branches had only one manager andrepresented branches of smaller sizes. We tested the hypotheses with andwithout the data from these 9 branches and found the pattern of the resultsto be highly consistent. We retained these branches in our formal analysesin order to have a more representative sample of the population.


Kozlowski & Hattrup, 1992). Therefore, we proceeded withaggregation, acknowledging that the relationships between themanagement-HPWS and the other study variables may be un-derestimated.

Measures of Variables Relevant for PredictingCustomer Satisfaction

Individual-level customer overall satisfaction with branch ser-vice. Branch customers provided evaluations of their satisfactionwith the branch’s overall service using eight items to assess the keyindicators of service quality for the Japanese banking industry (e.g.,teller performance, visiting service provided to the customer’shome or workplace). Alpha for this scale was .88.

Branch-level average employee service performance. In orderto examine how customer satisfaction is influenced by the overallservice performance of the branch employees, we aggregatedemployee ratings of the branch’s overall general service perfor-mance and knowledge-intensive financial service performance tothe branch level to represent the branch’s employee overall generalservice performance and knowledge-intensive service perfor-mance. Alpha was .96 for the former and .96 for the latter. Theaggregation was justified by a high median rwg(j) of .89 and .92,respectively, for the two types of service performance.

Control Variables

Individual-level control variables. We controlled for employ-ees’ age and gender at the individual employee level of analysis;information on these variables was obtained from the headquar-ters’ archival data. In addition, in assessing the effect of employeepsychological empowerment or intrinsic motivation on employeeservice performance, we controlled for the influence of extrinsicmotivation. We measured employee-perceived extrinsic motiva-tion following the procedure described in Tyagi (1985): First,performance expectancy (E; 6 items; alpha � .92), as well as theinstrumentality (I; 5 items; alpha � .84) and valence (V; 5 items;alpha � .83) of extrinsic rewards for performance, were measuredwith items answered by employees adapted from Gwinner et al.(2005); then, the extrinsic motivation score (EM) was calculatedby standardizing the score of

EM � �E � � �k �1

n

Ik � Vk��.

When predicting a customer’s satisfaction with the branch’soverall service, we controlled for the customer’s gender, age, andtotal household asset as reported by the responding customer at theindividual customer level of analysis.

Group-level control variables. We controlled for employees’affiliation with a specific employment group at the group level,using the archival data from the bank’s headquarters (Group 1 �1 if the employee belonged to employee Group 1; Group 2 � 1 ifthe employee belonged to employee Group 2; and Group 3 wasused as the comparison group).

In addition, we controlled for the average ratings of employee-HPWS at the group level, which represents the employees’ sharedexperience of the HPWS within the same group in a branch.Through social interactions, employees of the same group may

come to form certain common perceptions or psychological states,which may further influence employee attitudes and behaviorabove and beyond employee individual experience with the worksystem (e.g., Liao & Chuang, 2007; Ostroff & Bowen, 2000).Therefore, we aggregated employee-HPWS to the group level as acontrol variable in order to show that, after accounting foremployee-shared experience of HPWS, an individual employee’sidiosyncratic experience with the HPWS (measured as a Level 1variable, employee-HPWS) was significantly related to employeehuman capital, psychological empowerment, perceived organiza-tional support, and performance. Aggregation is justified by a highrwg(j) adjusted for a small negative skew in the expected variance(James et al., 1984); the median rwg(j) value was .96.

Branch-level control variables. At the branch level, we con-trolled for the branch’s age, size, and number of competing banksin the neighborhood in all of the analyses. Information for thesevariables was provided by the bank’s headquarters. In addition,when testing Hypothesis 5 on the relationship between a branch’soverall employee service performance and customer satisfaction,we controlled for branch average management-HPWS (calculatedby first assigning the management-rated HPWS score for a partic-ular employee group to all of the employees in that group, thenaveraging the scores across all the employees in the branch),branch average employee-HPWS, and branch average employeehuman capital, psychological empowerment, extrinsic motivation,and perceived organizational support. We do not assume a highagreement across individuals on these variables within a branch;they are controlled for merely to demonstrate the effect of branchoverall employee service performance on customer satisfactionbeyond the potential effects of these variables on customer satis-faction at the branch level.

Analytical Strategy

In this study, employees were nested in employment groups, andemployment groups were nested in branches. The theoretical mod-els involving employee outcomes are also hierarchical, with con-structs spanning three levels of analysis. Specifically, employeeindividual service performance, employee-HPWS, human capital,psychological empowerment, and perceived organizational sup-port are conceptualized at the individual employee level of anal-ysis; management-HPWS targeted at three different groups withinthe branch is conceptualized at the group level of analysis; andbranch characteristics, such as size, age, and level of local com-petition, are conceptualized at the branch-level of analysis. Inaddition, customers are nested in branches. The theoretical modelsinvolving customer satisfaction as the outcome also have twolevels, with individual customers’ satisfaction and demographiccharacteristics conceptualized at the individual customer level ofanalysis and branch collective service performance conceptualizedat the branch level of analysis.

Therefore, we conducted three-level hierarchical linear model-ing (HLM; Bryk & Raudenbush, 1992) to test Hypotheses 1through 4, involving employee outcomes, and two-level HLM totest Hypothesis 5, involving customer outcomes. Before conduct-ing the HLM analyses, we conducted a series of analyses ofcovariance (ANCOVAs), analyses of variance (ANOVAs), andgroup mean comparison tests to examine differences in employeeexperience with the work system across different groups, the


variance within the same group, and the divergence betweenmanager and employee perspectives of the work system. Theseanalyses, although not part of our formal hypothesis testing, serveto provide the empirical examination for the potential importanceof focusing on employee individually experienced HPWS.

Results

Descriptive statistics, correlations, and alpha values for thestudy variables are presented in Table 2.

Employee–High Performance Work System:Differences Between Employee Groups, VarianceWithin a Group, and Divergence From theManagement’s Perspective

Differences between employee groups. In order to understandwhether there are significant differences in employee-HPWSamong the three different employee groups, we conductedANCOVA at the individual level, with group membership as thetreatment variable and employee gender and age as the covariates.The results revealed that group membership was a significantfactor in determining employee-HPWS, F(2, 826) � 4.84, p �.0001. We further conducted pairwise mean comparisons for em-ployee Group 1, Group 2, and Group 3 across branches. The resultsrevealed significant differences between Group 1 and Group 2 (forGroup 1, M � 3.49, SD � 0.02; for Group 2, M � 3.19, SD �0.03; t(87) � 10.68, p � .001, n � 88), between Group 2 andGroup 3 (for Group 2, M � 3.21, SD � 0.03; for Group 3, M �3.03, SD � 0.04; t(70) � 3.80, p � .001, n � 71), and betweenGroup 1 and Group 3 (for Group 1, M � 3.49, SD � 0.02; forGroup 3, M � 3.05, SD � 0.04; t(73) � 10.68, p � .001, n � 74).3

Across branches, Group 1 generally had the most favorable eval-uations of the work system, followed by Group 2, and then Group3, which is consistent with the company’s categorization of em-ployment status.

Variance within a group. We also examined whether therewere significant between-employee differences in participants’experience of HPWS within employment status groups. Specifi-cally, we followed the procedure recommended in Bryk and Rau-denbush (1992) and performed an ANOVA using a two-level nullHLM analysis (for Level 1, n � 831; for Level 2, n � 253), withemployee individual experience of the work system as the outcomevariable. The results showed that the within-group variance esti-mate was .15 and the between-group variance estimate was .03( p � .001). This indicates that 17% of the variance in employeeexperience of HPWS resided between employee groups, whereas83% of the variance resided within employee groups.

Divergence from the management’s perspective. Our dataalso allowed us to examine whether for a given employmentgroup, there was a discrepancy between employee-HPWS andmanagement-HPWS. First, we performed mean comparisonsbetween management and employee perspectives of the worksystem. The results revealed that across the 253 employeegroups in different branches, the group mean of management-HPWS for a certain group was 3.50 (SD � 0.02), and the groupmean of employee-HPWS for the same group was 3.25 (SD �0.02), which was significantly different from, and lower than,the management-HPWS, t(252) � �10.58, p � .001, n � 253.

This suggests that management generally tends to have a morepositive evaluation of the work system than do the employees.On the other hand, the correlation at the group-level betweenmanagement-HPWS for a group and the average employee-HPWS in that group was .39, and its 95% confidence intervaldoes not include 0 or 1. These results suggest that overall thereis a divergence between management-HPWS and the groupmean of employee-HPWS, although the two perspectives arepositively related to each other.

Three-Level Hierarchical Linear Modeling PredictingIndividual Employee Outcomes

Hypotheses 1 through 3 predicted that employee-HPWS wouldpositively influence employee service performance through themediation of employee human capital, psychological empower-ment, and perceived organizational support. We followed thefour-step test procedures for mediation described in Kenny, Kashy,and Bolger (1998); controlled for the effects of management-HPWS and a variety of individual, group, and branch characteris-tics; and report the results in Tables 3 and 4. As a first step,employee-HPWS needs to be related to employee service perfor-mance. As shown in Model 2 and Model 5 in Table 4, employee-HPWS had a significant positive relationship with supervisor-ratedemployee general service performance (̂ � .48, p � .001) andknowledge-intensive service performance (̂ � .24, p � .001).

To test Hypothesis 1 regarding the role of human capital as themediator, in the second step, we examined whether employee-HPWS was related to the mediator. As shown in Model 3 in Table3, employee-HPWS was positively related to human capital (̂ �.51, p � .001). Next, in testing the third and fourth steps, weincluded both employee-HPWS and the mediator in predictinggeneral service performance and knowledge-intensive service per-formance. As reported in Models 3 and 6 in Table 4, employee-HPWS was no longer significant, yet human capital was signifi-cantly related to general service performance (̂ � .70, p � .001)and knowledge-intensive service performance (̂ � .29, p � .001).Therefore, Hypothesis 1 was supported.

We followed the same steps in testing Hypothesis 2 regardingthe role of psychological empowerment as a mediator. As reportedin Model 5 of Table 3, employee-HPWS was positively related toempowerment (̂ � .53, p � .001). Models 3 and 6 in Table 4showed that empowerment was not significantly related to generalservice performance but was significantly related to knowledge-intensive service performance (̂ � .14, p � .01). Therefore,empowerment mediated the relationship between employee-HPWS and knowledge-intensive service performance, partiallysupporting Hypothesis 2.

Likewise, Hypothesis 3 was tested by first examining the relation-ship between employee-HPWS and the mediator, perceived organi-zational support, which was shown to be significant in Model 7 ofTable 3 (̂ � 1.43, p � .001). Next, in Models 3 and 6 of Table 4,whereas employee-HPWS became nonsignificant in predicting ser-vice performance, perceived organizational support was significantly

3 Because not all branches had data for all three groups, the group meansand sample size involved in the mean comparison tests were slightlydifferent.


Table 2Descriptive Statistics, Intercorrelations, and Internal Consistency of Study Variables

Variable M SD 1 2 3 4 5 6 7 8 9 10 11 12

Individual�level employee variables (N � 830)

1. Employee gendera 0.57 0.50 —2. Employee age 32.58 8.76 �.37� —3. Employee�HPWS 3.25 0.43 �.31� .05 .824. Supervisor�rated employee

human capital 4.52 1.20 .11� �.04 .15� .945. Employee psychological

empowerment 3.22 0.51 �.34� .26� .49� .27� .896. Employee extrinsic

motivation �0.02 1.00 �.45� .07� .42� .04 .48� —7. Employee POS 5.11 0.92 �.12� �.08� .65� .18� .27� .24� .898. Supervisor�rated employee

GSP 4.89 0.96 .04 .03 .20� .89� .29� .06 .23� .979. Supervisor�rated employee

KSP 3.14 0.93 �.63� .28� .33� .35� .42� .38� .19� .45� .96

Group�level employee variables (N � 252)

1. Employee Group 1 0.36 0.48 —2. Employee Group 2 0.35 0.48 �.55� —3. Employee Group 3 0.29 0.46 �.48� �.47� —4. Management�HPWS 3.50 0.36 .26� .30� �.59� —5. Group average

employee�HPWS 3.25 0.31 .55� �.15� �.32� .39� —

Individual�level customer variables (N � 1,772)

1. Customer gendera 0.54 0.50 —2. Customer age 48.83 13.58 .02 —3. Customer total household

asset 3.21 2.07 .08� .45� —4. Customer overall

satisfaction with branchservice 4.45 0.78 .05� .27� .17� .88

Branch�level variables (N � 75)

1. Branch age 40.11 18.39 —2. Branch size 15.56 7.36 .46� —3. Number of competing banks

in neighborhood 3.58 2.95 .05 .39� —4. Branch average

management�HPWS 3.55 0.24 .12 �.02 .00 —5. Branch average

employee�HPWS 3.33 0.14 �.02 �.17 .� 03 .29� —6. Branch average human

capital 4.48 0.58 .13 .18 �.12 .24� .18 —7. Branch average

psychological empowerment 3.38 0.22 �.16 �.36� �.13 .12 .44� .03 —8. Branch average extrinsic

motivation 0.05 0.33 .00 �.07 .09 .25� .49� �.16 .51� —9. Branch average POS 5.17 0.39 �.10 �.09 .00 .23� .73� .26� .24� .16 —

10. Branch average overall GSP 5.05 0.30 �.16 �.07 .01 .15 .56� .18 .49� .39� .53� —11. Branch average overall KSP 3.70 0.23 �.07 �.14 �.05 .04 .26� �.13 .39� .33� .16 .51� —12. Branch average customer

overall satisfaction 4.46 0.25 .12 �.14 �.36� .04 .03 �.02 .19 .10 .06 .18 .37� —

Note. Numbers 1–12 in the top row correspond to the variables in the respective sections of the table. Coefficient alpha values are presented in italicsalong the diagonal. Employee-HPWS � Employee individually experienced high performance work system; POS � perceived organizational support;GSP � general service performance; KSP � knowledge-intensive financial service performance; management-HPWS: Management’s perspective of theHPWS practices generally in use for an employee group of certain employment status. Branch average management-HPWS was calculated by averagingacross branch employees the management-HPWS score assigned to the employees.a Female � 1; male � 0.� p � .05.


related to general service performance (̂ � .06, p � .05) but was notsignificantly related to knowledge-intensive service performance.Therefore, Hypothesis 3 was partially supported, in that perceivedorganizational support mediated the relationship between employee-HPWS and general service performance.

We conducted Sobel (1982) tests and confirmed that the change inthe significance of employee-HPWS in predicting general serviceperformance due to the introduction of human capital (z � 4.09, p �.001) and perceived organizational support (z � 2.43, p � .05) wassignificant. Similarly, Sobel tests revealed that human capital (z �3.93, p � .001) and psychological empowerment (z � 2.93, p � .01)significantly reduced the significance of employee-HPWS in predict-ing knowledge-intensive service performance.

In sum, employee human capital and perceived organizationalsupport fully mediated the relationship between employee-HPWS andemployee general service performance, whereas human capital andpsychological empowerment fully mediated the relationship betweenemployee-HPWS and employee knowledge intensive service perfor-mance, providing some support for Hypotheses 1 through 3.

We followed a similar procedure to test Hypothesis 4, whichproposes that management-HPWS would influence employee humancapital, psychological empowerment, and perceived organizationalsupport through the mediation of employee-HPWS. As reported inModel 1 of Table 3, management-HPWS was not significantly relatedto employee-HPWS. Therefore Hypothesis 4 was not supported. Asshown in Models 3, 5, and 7 of Table 3, management-HPWS waspositively related to human capital (̂ � .56, p � .01) but wasinsignificantly related to psychological empowerment or perceivedorganizational support. The results suggest that management-HPWS

had a direct effect on individual employees’ human capital instead ofa mediated effect through employee-HPWS.

Two-Level Hierarchical Linear Modeling PredictingCustomer Satisfaction

Hypothesis 5 predicted that branch overall service performancewould be positively related to customer overall satisfaction withthe branch’s service. We conducted HLM analysis to test thishypothesis, treating customer satisfaction as an individual cus-tomer level outcome. This approach allowed us to achieve twogoals, which could not be achieved if we had aggregated customersatisfaction to the branch level and used collective performance topredict aggregated customer satisfaction using an ordinary leastsquares regression. First, we are able to account for customers’individual differences in age and gender, which have been shownto relate to customer satisfaction ratings (e.g., Liao & Chuang,2004). Second, the number of customer respondents per branchranged from 1 to 62, with an average of 24; HLM can account forthis differential precision of information of each branch by usinggeneralized least squares to estimate the fixed effects, which aretypically reported as the final estimates of the impact of thepredictors on the outcome variables (Hofmann, 1997). In the GLSprocedure, branches with more observations, hence, more reliableand precise Level 1 estimates, receive more weight (Bryk &Raudenbush, 1992). Fixed effects can thus be viewed as theweighted average of the Level 1 coefficients across branches. Theresults in Model 4 of Table 5 revealed that branch overallknowledge-intensive service performance was positively related to

Table 3Three-Level Hierarchical Linear Modeling Results for Employee High-Performance Work Systems and Mediators

Variable

Employee-HPWS

Employee knowledge,skills, and abilities:

Supervisor-ratedemployee human capital

Employee motivation:Employee psychological

empowerment

Employee obligation:Employee perceivedorganization support

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7

Intercept 3.43�� 4.13�� 3.99�� 3.19�� 3.04�� 5.73�� 5.35��

Level 1: Employee levelEmployee gendera �.33�� .74�� .89�� .16 .001 �.36 .04Employee age �.002 .004 .004 .01�� .01�� .02�� .01��

Employee-HPWS .51�� .53�� 1.43��

Level 2: Group levelEmployee Group 1b �.02 �.05 .13 .08 .27�� .77�� .29Employee Group 2b .08 �.27� �.29� .19�� .17� �.30�� .35��

Management-HPWS .09 .53�� .56�� .11 �.07 .03 .12Group average employee-HPWS .74�� .25 .49�� .04 1.45�� .02

Level 3: Branch levelBranch age �.001 .003 .003 �.001 �.001 �.003 �.003Branch size �.003 .01 .01 .000 .001 �.003 �.004Number of competing banks in

neighborhood �.001 �.01 �.01 .000 .000 �.001 .000Pseudo R2c 0.12 0.07 0.09 0.19 0.32 0.17 0.46

Note. For Level 1, N � 830; for Level 2, N � 252; for Level 3, N � 91. In all models, all variables except for employee gender, employee Group 1, andemployee Group 2 were grand-mean centered. Entries corresponding to the predicting variables are estimations of the fixed effects ( values) with robuststandard errors. Employee-HPWS � employee individually experienced HPWS; management-HPWS � management’s perspective of the HPWS practicesgenerally in use for an employee group of certain employment status.a Female � 1; male � 0. b The omitted, comparison employee group consisted of part-time employees. c Pseudo R2 values were calculated on the basisof the formula from Kreft and De Leeuw (1998).� p � .05. �� p � .01. �� p � .001.


customer overall satisfaction (̂ � .34, p � .05), providing somesupport for Hypothesis 5.

Discussion

The primary goal of this study was to examine how employee-HPWS affected individual performance in the service context.Although researchers in the area of strategic HRM research havesuggested that management and employee perspectives might di-verge (e.g., Bowen & Ostroff, 2004; Wright & Boswell, 2002), thecurrent study is among the first to examine the difference andrelationship between the management and employee perspectivesof HPWS. In addition, in the strategic HRM literature, researchershave examined the link between the use of HPWS and a myriad oforganizational performance outcomes at the firm level of analysis.However, organizations do not “perform”; individuals in organi-zations perform in ways that allow the organizations to achievedesirable performance outcomes (Kozlowski & Klein, 2000).Thus, individual performance remains an important performancecriterion for management and psychology research to assess theeffectiveness of work systems. The current study sheds light on theimpact and influence process of a service-quality–oriented HPWSon individual performance provided to customers. The resultsprovide several implications for research and practice.

Research Implications

One of the major implications of the findings for this study isthat there is a disconnect between what management says they are

implementing and what employees report they are experiencing interms of the HPWS practices. We found that although the corre-lation between management-HPWS and group average employee-HPWS was positive overall, the managerial ratings were signifi-cantly higher than employee ratings of the HPWS. In addition, theeffect of management-HPWS on individual employee-HPWS wasnot significant when they were entered into the three-level hierar-chical linear model. Relatedly, our analyses revealed substantialvariance in employee-HPWS among groups of employees withdifferent status and among employees within the same group.These findings demonstrate the variability in HR practices acrossemployees within the same unit, which has been largely ignored byprior strategic HRM research (Wright & Boswell, 2002). Thefindings thus provide empirical support for the arguments of Lepakand Snell (1999, 2002) that multiple work systems may existwithin the same establishment and for the arguments of Guzzo andNoonan (1994) as well as Bowen and Ostroff (2004) that individ-ual employees may experience and interpret the same set of HRpractices differently.

A second major implication is that the results highlight areas ofdistinction and overlap between the impact of management perspec-tive of the HPWS practices generally implemented and employeeindividual experience with the HPWS on individual employee out-comes. From the employee perspective, employee-HPWS had a directpositive impact on employee human capital, psychological empow-erment, and perceived organizational support, which were in turnrelated to general and knowledge-intensive service performance. Hu-man capital and perceived organizational support fully mediated the

Table 4Three-Level Hierarchical Linear Modeling Results Predicting Individual Employees’ Service Performance

Level and variable

Supervisor-rated employee generalservice performance

Supervisor-rated employee knowledge-intensive financial service performance

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6

Intercept 4.59�� 4.45�� 4.83�� 3.10�� 3.03�� 3.24��

Level 1: Individual levelEmployee gendera .55�� .70�� .05 �.38�� .31� �.57��

Employee age .01 .01 .01�� .01�� .01�� .01��

Employee-HPWS .48�� .06 .24�� .01Supervisor-rated employee human capital .70�� .29��

Employee psychological empowerment .003 .14��

Employee extrinsic motivation �.02 .02Employee perceived organizational support .06� .01

Level 2: Group levelEmployee Group 1b �.02 .15 .06 .48�� .57�� .46��

Employee Group 2b �.20 �.22 �.01 �.005 �.01 .03Management-HPWS .40�� .42�� .03 .33� .35� .19Group average employee-HPWS .64�� .17 .02 .55�� .32� .28�

Level 3: Branch levelBranch age .002 .002 .000 .002 .002 .001Branch size .01 .01 .000 �.002 �.002 �.005Number of competing banks in neighborhood �.01 �.01 .001 �.01�� .01�� .01�

Pseudo R2c 0.06 0.09 0.80 0.44 0.45 0.60

Note. For Level 1, N � 830; for Level 2, N � 252; and for Level 3, N � 91. In all models, all variables except for employee gender, employee Group1, and employee Group 2 were grand-mean centered. Entries corresponding to the predicting variables are estimations of the fixed effects ( values), withrobust standard errors. HPWS � high-performance work system.a Female � 1; male � 0. b The omitted, comparison employee group consisted of Group 3, or part-time, employees. c Pseudo R2 values were calculatedon the basis of the formula from Kreft and De Leeuw (1998).� p � .05. �� p � .01. �� p � .001.


relationship between employee-HPWS and general service perfor-mance, whereas human capital and psychological empowerment fullymediated the relationship between employee-HPWS and knowledge-intensive service performance. These results provide some support forthe individual-level theory of job performance (Campbell et al.,1993), suggesting that individuals’ knowledge, skills, and abilities andmotivation are proximal determinants of individual performance andthat the employee experience with HR interventions influences jobperformance through its impact on the individuals’ knowledge, skills,and abilities and motivation. More importantly, the effects ofemployee-HPWS on these individual outcomes sustained after weaccounted for the effects of management-HPWS and employee aver-age perceptions of the HPWS within the group, demonstrating theutility of considering employees’ idiosyncratic experience with theHPWS above and beyond management perspective of the HR prac-tices generally implemented for an employee group and the employ-ees’ shared perceptions of the HR practices within the group.

Whereas employee-HPWS was directly related to both em-ployee human capital and motivation (psychological empower-ment and perceived organizational support), we found thatmanagement-HPWS was directly related only to employee humancapital. One possible explanation of this finding is that someaspects of employee attributes (i.e., motivation) are more directlyinfluenced by employee interpretation of the work system than areother aspects of employee attributes (i.e., human capital), and thusmanagement-HPWS may have a stronger direct impact on thelatter form of attributes. For example, providing training programsmay positively influence employee human capital independent ofemployees’ perceptions of the systems. If the management investsin the development of employee competencies for service delivery,their human capital—their knowledge, skills, and abilities to pro-

vide quality service—should be positively affected, even if theyare not aware of or do not perceive those investments as occurring.In contrast, influencing employee motivation to perform is depen-dent on employee personal understanding and interpretation of theHPWS practices.

Taken in combination, these results suggest that we should notassume homogeneity of employee experience with the HPWSacross employees of different employment groups or even withinthe same group, nor should we assume an equivalence of theemployee perspective with the management perspective. Thus, it isimportant to directly assess employees’ individual experienceswith the work system in theoretical development and empiricaltesting of the effects of the HPWS on individual-level employeeoutcomes. Future research may also examine explicitly what givesrise to employees’ differential experiences by considering thepotential antecedents we mentioned in this study, such as employ-ment status, organizational justice, and leader–member exchangerelationships. In addition, Bowen and Ostroff (2004) proposed aset of HR system meta-features that may influence the level towhich employees build shared perceptions about the HR system,including visibility of the HR practices, understandability of theHR content, legitimate authority of the HR system, relevance ofthe HR system to the strategic goal, instrumentality of the HRsystem for employee consequences, validity of the HR practices,consistency of the HR messages, agreement among principal HRdecision makers, and fairness of the practices. A concerted effortto assess the effects of these features may represent a promisingavenue for future research.

Beyond examining the divergence between, and the relativeimpact of, management-HPWS and employee-HPWS on employeehuman capital and motivation and, in turn, on service performance,

Table 5Three-Level Hierarchical Linear Modeling Results Predicting Individual Customers’ OverallSatisfaction With Branch Service

Level and variable Model 1 Model 2 Model 3 Model 4

Intercept 4.41�� 4.41�� 4.42�� 4.42��

Level 1: Individual customer levelCustomer gendera .07 .07 .07 .07Customer age .01�� .01�� .01�� .01��

Customer total household asset .02 .02 .02 .02Level 2: Branch level

Branch age .000 .000 .001 .000Branch size �.001 �.001 �.001 �.001Number of Competing Banks in Neighborhood �.01 �.01 �.01 �.01Branch average management-HPWS �.05 �.06 �.06 �.03Branch average employee-HPWS .11 �.18 �.27Branch average human capital �.01 .01Branch average psychological empowerment .07 .02Branch average extrinsic motivation .03 �.01Branch average POS .14 .13Branch average overall GSP .004Branch average overall KSP .34��

Pseudo R2 0.071 0.071 0.069 0.074

Note. For Level 1, N � 1,772; for Level 2, N � 75. In all models, Level 1 variables except for customer gender weregrand-mean centered. Entries corresponding to the predicting variables are estimations of the fixed effects ( values),with robust standard errors. POS � perceived organizational support; GSP � general service performance; KSP �knowledge-intensive financial service performance.a Female � 1; male � 0.�� p � .01. �� p � .001.


we also examined the linkages between branch-level service per-formance and customer satisfaction. Of the two types of serviceperformance, employees’ overall knowledge-intensive service per-formance at the branch level was positively related to customersatisfaction with branch overall service quality, whereas employeegeneral service performance was not. This result suggests that allperformance metrics might not be created equal in helping firmsgain a competitive service advantage. The resource-based view ofthe firm (Barney, 1991) suggests that sustainable competitiveadvantage comes from a firm’s resources and capabilities that arevaluable, rare, imperfectly imitable, and not substitutable. Wesurmise that the behaviors of general service performance, such asbeing courteous, attentive, and reliable to customers and having aneat, professional appearance may be strictly scripted, or required,behaviors in banking services and part of customers’ standardexpectation of a service encounter; thus, it may be difficult forbranches to compete for higher customer satisfaction on the basisof general service performance. Instead, superior knowledge-intensive service performance may be rarer and less imitable andhence help differentiate superior banking experiences from othersand create a competitive advantage in service. We encouragefuture researchers examining HPWS to keep in mind what perfor-mance they are referring to in the specific context and to gobeyond general performance metrics by paying attention to thespecial metrics that differentiate firms from their competitors.

We found that the HPWS, from both the management and theemployee perspectives, did not have a significant relationship withcustomer satisfaction. The lack of a direct link between the two,however, does not imply that the HPWS is not important incustomer service. The impact of the HPWS on customer satisfac-tion may be a distal effect, which is likely to be “transmittedthrough additional links in a causal chain” (Shrout & Bolger, 2002,p. 429). Schneider, Ehrhart, Mayer, Saltz, and Niles-Jolly (2005)also argued that researchers need not necessarily despair when keyorganizational variables do not seem to be related to the bottomline when investigating bivariate relationships. These authorscalled for the development of richer models with process variables,which may reveal that these key variables are indeed important,although not directly so. The current study does just that. Byintroducing employee service performance as the process variable,we bridge the gap between strategic HRM research and servicemanagement research. The indirect effect of the HPWS on cus-tomer satisfaction is consistent with the general argument of thestrategic HRM literature that HPWS affects employee behaviors,which, in turn, affect organizational operational and financialperformance (e.g., see Lepak et al., 2006; Ostroff & Bowen, 2000;Wright, Gardner, & Moynihan, 2003). This finding is also consis-tent with the growing evidence from the service linkage research,which suggests that through front-line employees’ service behav-iors, internal organizational management transforms into desirableexternal customer outcomes (e.g., Heskett, Sasser, & Schlesinger,1997; Liao, 2007; Liao & Chuang, 2004, 2007; Schneider et al.,1998; 2005).

Though not the main focus of this study, the results also pro-vided some interesting implications regarding gender differencesin service performance. We found that women were rated higherthan men in general service performance, whereas the reverse wastrue for knowledge-intensive service performance. This finding isconsistent with several lines of gender research. First, gender and

personality theory suggests that men and women are predisposedto different orientations, with men being more instrumental andwomen being more expressive (Evans & Steptoe, 2003). Ourmeasures of general service performance reflected more social,relational performance (i.e., reliability, responsiveness, assurance,and empathy), whereas knowledge-intensive service performanceindicated technical performance. One interpretation of this findingis that women may excel in social relations, whereas men mayexcel in their technical performance. Second, gender role theoryposits that because social norms prescribe women to be communaland men to be agentic, different genders are “prescribed” withdifferent behaviors and will try to match themselves with the socialnorm (Heilman & Chen, 2005). This reinforces the tendency forwomen to perform well in social aspects of service and for men toperform well in the technical aspect of service. Alternatively, fromthe observers’ perspective, gender stereotypes analogous to thefirst two points may influence the way supervisors evaluate sub-ordinates’ performance (Heilman & Haynes, 2005). Thus, men andwomen may be evaluated differently by supervisors who arepredisposed to emphasize certain aspects of their performance.Therefore, the gender differences we observed here might not be aunique phenomenon in Japan. Nonetheless, research is needed toexamine potential cultural influences in the role of gender onservice performance.

Lastly, we found that employees in Group 1 had betterknowledge-intensive financial service performance than did em-ployees in Group 3 or Group 2 after a variety of antecedents ofemployee performance had been accounted for. We surmise thatbecause Group 1 has the highest employment status, better per-formers are likely to be attracted to, selected by, and retained inthis group (attraction–selection–attrition [ASA] theory; Schneider,1987; Schneider, in press; Schneider, Goldstein, & Smith, 1995).Therefore, there seem to be other factors (e.g., availability ofresources) that make Group 1 employees better performers beyondall of the individual, group, and branch factors that we havespecified in this study. In any case, because we explicitly con-trolled for group membership in our analyses, we partialed out theeffects of the unspecified factors associated with group member-ship on performance.

Practical Implications

The current study also offers several practical implications.Marketing research has shown that customer satisfaction is asso-ciated with customer intention to repurchase and spread positiveword of mouth (e.g., Parasuraman, Zeithaml, & Berry, 1988) andwith financial measures such as return on assets and stock prices(e.g., Ittner & Larcker, 1996). For example, as Liao and Sub-ramony (2008) noted, marketing research has shown that a 1%increase in the American Customer Satisfaction Index (ACSI)score (Fornell, Johnson, Anderson, Cha, & Bryant, 1996) increasesa medium-sized firm’s (e.g., with $54 billion in assets) future cashflow by $55 million and reduces the variability of the firm’s cashflow by more than 4% (Gruca & Rego, 2005). Thus, customersatisfaction can help firms increase both the volume and thestability of their future cash flow, hence creating greater share-holder value. Our findings suggest that a key factor related tocustomer satisfaction in the banking service context is branchemployees’ knowledge-intensive service performance; the two


variables had a correlation of .37, with a 95% confidence intervalranging from .16 to .55, indicating that branch employeeknowledge-intensive service performance accounted for between3% and 30% of the variance in customer satisfaction. Thus, bankmanagement may achieve desirable customer and organizationalfinancial outcomes by providing high-quality professional service.Our results also provide concrete suggestions on how to improveknowledge-intensive service performance. By implementing a sys-tem of service-quality–oriented HPWS practices, employees mayacquire the knowledge, skills, and abilities relevant for the deliveryof superior professional service, become psychologically empow-ered to appreciate the significant meaning in the tasks, feel thecompetence and control they have in performing the tasks, and seethe impact they can make on organizational success.

In addition, we found that branch employees’ general serviceperformance was not significantly related to customer satisfaction.However, this does not suggest that general service performance isunimportant. In fact, a deviation from these typically standard andexpected behaviors will result in customer dissatisfaction(Solomon, Surprenant, Czepiel, & Gutman, 1985). Thus, compa-nies should not focus exclusively on employee knowledge-intensive performance, ignoring the necessity of employee generalservice performance. Our results suggest that implementing aservice quality-oriented HPWS will likely help maintain a highlevel of general service performance. Under such a system, em-ployees will acquire the human capital needed for general serviceperformance and perceive a high level of organizational supportwhich, in turn, motivates them to provide superior generalservice performance to customers as reciprocation to the orga-nization’s favorable treatment.

These findings also suggest that management should pay par-ticular attention to employees’ idiosyncratic experiences of thework system, which directly influence their level of human capital,psychological empowerment, perceived organizational support,and service performance. Because HR practices communicatemessages on what management expects, supports, and rewards(Bowen & Ostroff, 2004; Schneider, 1990), management needs toimprove communications and prevent well-intended HR practicesfrom being misunderstood by employees. Increased dialogue be-tween management and employees and regular use of employeesurveys and discussion groups may help management betterunderstand what employees actually experience in the work-place and reduce the discrepancy between management andemployee perspectives.

Limitations and Future Research

We acknowledge that the study findings should be considered inlight of several limitations. First, we conducted the analyses usingdata from establishments within a Japanese firm. As a result, theremay be some concerns regarding the generalizability of thesefindings to other cultural contexts. At the same time, however, theconceptual arguments used to derive the hypotheses are not cul-turally bound, and the findings are consistent with the conceptualarguments developed throughout the strategic HRM literature. Inaddition, we explicitly controlled for factors such as employeeemployment status, the categorization of which is unique in theJapanese culture context, making the results about the key studyvariables interpretable in other culture contexts. Relatedly, this

sample was based on 91 establishments from a single bankingorganization. It is possible that there may be a corporate culturethat constrains the choice of HR practices used in the establish-ments and employees’ reactions to those practices. Thus, rangerestriction might have rendered the effects of the work system onemployee and customer outcomes to be underestimated. At thesame time, however, the compatibility across branches in terms ofproducts, services, pricing, marketing strategies, and so forth helpsreduce concerns about potential confounding effects due to differ-ences on these factors. Nevertheless, future research needs to testthe relationships across a wider array of organizations.

A second potential limitation of this study is that we focusedonly on an HPWS targeted toward service quality. In reality,organizations need to juggle multiple demands and achieve mul-tiple goals simultaneously. For example, in the banking industry,service quality is a critical component of organizations’ success;meanwhile, organizations need to improve safety, reduce work-place discrimination and inequity, increase shareholder value, andact as responsible member of the community. Future theoreticaland empirical work is needed to provide guidance on how tobalance the needs and align the interests of different organizationalstakeholders in designing the work system.

A third limitation is that although we obtained data from fiverating sources in three time periods, because of practical con-straints, we were not able to completely eliminate common methodbias. In particular, employees provided evaluations of both theirown experiences with the work system and their psychologicalempowerment and perceived organizational support, rendering theobserved relationships among these variables subject to commonsource bias. However, conceptually it was necessary to haveemployees self-report these measures because it was the individ-uals’ idiosyncratic experience and perceptions that were of con-cern. In addition, we measured employee-HPWS and the otheremployee-rated variables at two separate times in two surveys,thus reducing the common method bias.

In addition, supervisors evaluated individual employees’ humancapital as well as their service performance. On the one hand, theperformance evaluation literature suggests that supervisory ratingsof the different aspects of a subordinate often are subject to halobias, which may explain the high correlation between the em-ployee human capital and general service performance measures.On the other hand, the high correlation may suggest that humancapital is a key driver of employee service performance (account-ing for 80% of the variance in employee general service perfor-mance). This is reasonable, as an employee with high levels ofservice-related knowledge, skills, and abilities should be in a betterposition to provide more reliable, responsive, and confident ser-vice. Nonetheless, we call for future research to examine therelationship using ratings from different sources. We should notethat even given the high correlation between human capital andemployee general service performance, employee perceived orga-nizational support still had a significant positive relationship withemployee general service performance beyond the effect of humancapital, providing convincing evidence for the importance of em-ployee perceived organizational support, thus lending support tothe social exchange theory.

A fourth limitation is the short time period under which thisresearch took place. Although the data were collected at differentpoints in time, it is possible that the effects of the work system on


individual employees may take longer to materialize. Longitudinalresearch that examines the relationship between HPWSs and theimportant outcomes over different lengths of time may provideunique insights into not only the nature of the relationships but alsothe time lag necessary to realize the benefits of the work systems.In addition, future research that adopts the field experiment meth-odology is especially needed to test the causal influences of thework systems on employee and organizational outcomes.

Lastly, we took a cross-level approach in the current study to fill thevoid in strategic HRM research by examining the effects and influ-ence process of the macrolevel HPWS on microlevel employeeperformance outcomes. Future research, similar to that reported inChen, Thomas, and Wallace (2005); DeShon, Kozlowski, Schmidt,Milner, and Weichmann (2004); and Ployhart, Weekley, andBaughman (2006), may model the exemplar multilevel research toexamine simultaneously the impact and influence processes ofHPWSs on performance outcomes at both the individual and unitlevels of analysis and to test the homology (e.g., Chen, Bliese, &Mathieu, 2005) for the effects of such systems at multiple levels.

Concluding Remarks

Despite its limitations, the current study has a number ofstrengths. The key theoretical advancements include the following:(a) examining the differences and relationship between manage-ment and employee perspectives of the HPWS for service quality,(b) integrating various theoretical perspectives to propose a cross-level framework for the effects of HPWS on individual-levelservice performance, and (c) directly testing, rather than assuming,the intermediate mechanisms. The key empirical and methodolog-ical strength lies in the use of HLM to account for the hierarchicalnature of our models and data and in the collection of data in twoformats from three phases and five rating sources. Our studydesign reduced common-method bias and allowed us to examinesimultaneously the interfaces between management, front-line em-ployees, and customers and to examine how internal managementpractices and employee psychological processes influence impor-tant external performance criteria. The key practical implicationsrelate to showing the greater importance of branch overallknowledge-intensive service performance over general serviceperformance in influencing customer satisfaction and to dem-onstrating that the HPWS for service quality was directly, aswell as indirectly, related to employee knowledge-intensiveservice performance.

In conclusion, this study bridges the gap between macroleveland microlevel approaches to HRM and extends strategic HRMresearch to the service arena and to individual level processes andperformance outcomes. The findings contribute to our understand-ing of how a service-quality–oriented HPWS influences individualservice performance, in part, through the mediation of employeehuman capital, psychological empowerment, and perceived orga-nization support, and highlight the importance of incorporating theemployee perspective into the examination of HPWS.

References

Allen, D., Shore, L. M., & Griffeth, R. W. (2003). The role of perceivedorganizational support and supportive human resource practices in theturnover process. Journal of Management, 29, 99–118.

Anderson, J. C., & Gerbing, D. W. (1988). Structural equation modeling inpractice: A review and recommended two-step approach. PsychologicalBulletin, 103, 411–423.

Asubonteng, P., & McCleary, K. J. (1996). SERVQUAL revisited: Acritical review of service quality. Journal of Services Marketing, 10,62–81.

Bagozzi, R. P., Yi, Y., & Phillips, L. W. (1991). Assessing constructvalidity in organizational research. Administrative Science Quarterly,36, 421–458.

Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioralchange. Psychological Review, 84, 191–215.

Barney, J. B. (1991). Firm resources and sustained competitive advantage.Journal of Management, 17, 99–120.

Batt, R. (2002). Managing customer services: Human resources practices,quit rates, and sales growth. Academy of Management Journal, 45,587–597.

Becker, B., & Gerhart, B. (1996). The impact of human resource manage-ment on organizational performance: Progress and prospects. Academyof Management Journal, 39, 779–801.

Becker, B. E., & Huselid, M. A. (1998). High performance work systemsand firm performance: A synthesis of research and managerial implica-tions. In G. R. Ferris (Ed.), Research in personnel and human resourcesmanagement (Vol. 16, pp. 53–101). Greenwich, CT: JAI Press.

Blau, P. M. (1964). Exchange and power in social life. New York: Wiley.Bliese, P. D. (2000). An introduction to multilevel modeling techniques

[Book review]. Personnel Psychology, 53, 1062–1065.Bowen, D. E., & Ostroff, C. (2004). Understanding HRM-firm perfor-

mance linkages: The role of the “strength” of the HRM system. Academyof Management Review, 29, 203–221.

Bowen, D. E., & Schneider, B. (1988). Services marketing and manage-ment: Implications for organizational behavior. Research in Organiza-tional Behavior, 10, 43–80.

Bryk, A. S., & Raudenbush, S. W. (1992). Hierarchical linear models.Newbury Park, CA: Sage.

Campbell, J. P., McCloy, R. A., Oppler, S. H., & Sager, C. E. (1993). Atheory of performance. In N. Schmitt, & W. C. Borman (Eds.), Person-nel selection in organizations: 35–70. San Francisco: Jossey-Bass.

Chen, G., & Bliese, P. D. (2002). The role of different levels of leadershipin predicting self- and collective efficacy: Evidence for discontinuity.Journal of Applied Psychology, 87, 549–556.

Chen, G., Bliese, P. D., & Mathieu, J. E. (2005). Conceptual frameworkand statistical procedures for delineating and testing multilevel theoriesof homology. Organizational Research Methods, 8, 375–409.

Chen, G., Thomas, B., & Wallace, J. C. (2005). A multilevel examinationof the relationships among training outcomes, mediating regulatoryprocesses, and adaptive performance. Journal of Applied Psychology,90, 827–841.

Colquitt, J. A., Conlon, D. E., Wesson, M. J., Porter, C. O., & Yee, N. K.(2001). Justice at the millennium: A meta-analytic review of 25 years oforganizational justice research. Journal of Applied Psychology, 86, 425–445.

Datta, D. K., Guthrie, J. P., & Wright, P. M. (2005). Human resourcemanagement and labor productivity: Does industry matter? Academy ofManagement Journal, 48, 135–145.

Deci, E. L. (1975). Intrinsic motivation. New York: Plenum.Delery, J. E. (1998). Issues of fit in strategic human resource management:

Implications for research. Human Resource Management Review, 8,289–309.

Delery, J. E., & Doty, D. H. (1996). Modes of theorizing in strategic humanresource management: Tests of universalistic, contingency, and config-urational performance predictions. Academy of Management Journal,39, 802–835.

Delery, J. E., & Shaw, J. D. (2001). The strategic management of peoplein work organizations: Review, synthesis, and extension. In G. R. Ferris


(Ed.), Research in personnel and human resources management (Vol.20, 165–197). Stamford, CT: JAI Press.

DeShon, R. P., Kozlowski, S. W. J., Schmidt, A. M., Milner, K. R., &Weichmann, D. (2004). A multiple goal, multilevel model of feedbackeffects on the regulation of individual and team performance in training.Journal of Applied Psychology, 89, 1035–1056.

Drasgow, F., & Kanfer, R. (1985). Equivalence of psychological measure-ment in heterogeneous populations. Journal of Applied Psychology, 70,662–680.

Eisenberger, R., Cummings, J., Armeli, S., & Lynch, P. (1997). Perceivedorganizational support, discretionary treatment, and job satisfaction.Journal of Applied Psychology, 82, 812–820.

Eisenberger, R., Fasolo, P., & Davis-LaMastro, V. (1990). Perceivedorganizational support and employee diligence, commitment, and inno-vation. Journal of Applied Psychology, 75, 51–59.

Evans, O., & Steptoe, A. (2003). Gender-related psychological character-istics and situational determinants of psychophysiological stress reactiv-ity. Journal of Applied Social Psychology, 33, 756–774.

Fornell, C., Johnson, M. D., Anderson, E. W., Cha, J., & Bryant, B. E.(1996). The American Customer Satisfaction Index: Nature, purpose,and findings. Journal of Marketing, 58, 7–18.

Graen, G. B., & Uhl-Bien, M. (1995). Relationship-based approach toleadership: Development of leader–member exchange (LMX) theory ofleadership over 25 years: Applying a multi-level multi-domain perspec-tive. Leadership Quarterly, 6, 219–247.

Gruca, T. S., & Rego, L. L. (2005). Customer satisfaction, cash flow, andshareholder value. Journal of Marketing, 69, 115–130.

Gutek, B. A. (1995). The dynamics of service: Reflections on the changingnature of customer/provider interactions. San Francisco, CA: Jossey-Bass.

Gutek, B. A., Bhappu, A. D., Liao-Troth, M. A., & Cherry, B. (1999).Distinguishing between service relationships and encounters. Journal ofApplied Psychology, 84, 218–233.

Guthrie, J. P. (2001). High-involvement work practices, turnovers, andproductivity: Evidence from New Zealand. Academy of ManagementJournal, 44, 180–190.

Guzzo, R. A., & Noonan, K. A. (1994). Human resource practices ascommunications and the psychological contract. Human Resource Man-agement, 33, 447–462.

Gwinner, K. P., Bitner, M. J., Brown, S. W., & Kumar, A. (2005). Servicecustomization through employee adaptiveness. Journal of Service Re-search, 8, 131–148.

Heilman, M., & Chen, J. (2005). Same behavior, different consequences:Reactions to men’s and women’s altruistic citizenship behavior. Journalof Applied Psychology, 90, 431–441.

Heilman, M. E., & Haynes, M. C. (2005). No credit where credit is due:Attributional rationalization of women’s success in male–female teams.Journal of Applied Psychology, 90, 905–916.

Heskett, J., Sasser, W. E. J., & Schlesinger, L. (1997). The service profitchain: How leading companies link profit and growth to loyalty, satis-faction, and value. New York: Free Press.

Hofmann, D. A. (1997). An overview of the logic and rationale of hierar-chical linear models. Journal of Management, 23, 723–744.

Huselid, M. A. (1995). The impact of human resource management prac-tices on turnover, productivity, and corporate financial performance.Academy of Management Journal, 38, 635–672.

Ittner, C., & Larcker, D. F. (1996). Measuring the impact of qualityinitiatives on firm financial performance. In S. Ghosh & D. Fedor (Eds.),Advances in the management of organizational quality (pp. 1–37).Greenwich, CT: JAI.

Jackson, S. E., & Schuler, R. S. (1995). Understanding human resourcemanagement in the context of organizations and their environments.Annual Review of Psychology, 46, 237–264.

James, L. R., Demaree, R. G., & Wolf, G. (1984). Estimating within-group

interrater reliability with and without response bias. Journal of AppliedPsychology, 69, 85–98.

James, L. R., James, L. A., & Ashe, D. K. (1990). The measuring oforganizations: The role of cognition and values. In B. Schneider (Ed.),Organizational climate and culture (pp. 40–89). San Francisco: Jossey-Bass.

Joshi, A., Liao, H., & Jackson, S. E. (2006). Cross-level effects of work-place diversity on sales performance and pay. Academy of ManagementJournal, 49, 459–481.

Kanter, R. M. (1983). The change masters: Innovation and entrepreneur-ship in the American corporation. New York: Simon & Schuster.

Keltner, B. (1995). Relationship banking and competitive advantage: Ev-idence from the U.S. and Germany. California Management Review, 37,45–73.

Kenny, D. A., Kashy, D. A., & Bolger, N. (1998). Data analysis in socialpsychology. In D. Gilbert, S. Fiske & G. Lindzey (Eds.), The handbookof social psychology (4th ed., Vol. 1, pp. 233–265). Boston: McGraw-Hill.

Kishton, J. M., & Widaman, K. F. (1994). Unidimensional versus domainrepresentative parceling of questionnaire items: An empirical example.Educational and Psychological Measurement, 54, 757–765.

Kozlowski, S. W. J., & Hattrup, K. (1992). A disagreement about within-group agreement: Disentangling issues of consistency versus consensus.Journal of Applied Psychology, 77, 161–167.

Kozlowski, S. W. J., & Klein, K. J. (2000). A multilevel approach to theoryand research in organizations: Contextual, temporal, and emergent pro-cesses. In K. J. Klein, & S. W. J. Kozlowski (Eds.), Multilevel theory,research, and methods in organizations (pp. 3–90). San Francisco, CA:Jossey-Bass.

Kreft, I., De Leeuw, J. (1998). Introducing multilevel modeling. London:Sage.

Lado, A. A., & Wilson, M. C. (1994). Human resource systems andsustained competitive advantage: A competency-based perspective.Academy of Management Review, 19, 699–727.

Lambert, S. J. (2000). Added benefits: The link between work–life benefitsand organizational citizenship behavior. Academy of Management Jour-nal, 43, 801–815.

Lepak, D. P., Liao, H., Chung, Y., & Harden, E. (2006). A conceptualreview of human resource management systems in strategic humanresource management research. In J. Martocchio (Ed.), Research inpersonnel and human resource management (pp. 217–271). Stamford,CT: JAI Press.

Lepak, D. P., & Snell, S. A. (1999). The human resource architecture:Toward a theory of human capital allocation and development. Academyof Management Review, 24, 31–48.

Lepak, D. P., & Snell, S. A. (2002). Examining the human resourcearchitecture: The relationships among human capital, employment, andhuman resource configurations. Journal of Management, 28, 517–543.

Lepak, D. P., Taylor, M. S., Tekleab, A. G., Marrone, J. A., & Cohen, D. J.(2007). An examination of the use of high-investment human resourcesystems for core and support employees. Human Resource Management,46, 223–246.

Levitt, T. (1972). Production line approach to services. Harvard BusinessReview, 50, 41–50.

Liao, H. (2007). Do it right this time: The role of employee servicerecovery performance in customer perceived justice and customer loy-alty after service failures. Journal of Applied Psychology, 92, 475–489.

Liao, H., & Chuang, A. (2004). A multilevel investigation of factorsinfluencing employee service performance and customer outcomes.Academy of Management Journal, 47, 41–58.

Liao, H., & Chuang, A. (2007). Transforming service employees andclimate: A multilevel, multisource examination of transformational lead-ership in building long-term service relationships. Journal of AppliedPsychology, 92, 1006–1019.


Liao, H., & Subramony, M. (2008). Employee customer orientation inmanufacturing organizations: Joint influences of customer proximity andsenior leadership team. Journal of Applied Psychology, 93, 317–328.

Liden, R. C., & Graen, G. B. (1980). Generalizability of the vertical dyadlinkage model of leadership. Academy of Management Journal, 23,451–465.

Melian-Gonzalez, S., & Verano-Tacoronte, D. (2006). Is there more thanone way to manage human resources in companies? Personnel Review,35, 29–50.

Miles, R. E., & Snow, C. C. (1984). Designing strategic human resourcessystems. Organizational Dynamics, (Summer), 36–52.

O’Farrell, P. N., Hitchens, D. M., & Moffat, L. A. R. (1993). The com-petitive advantage of business service firms: A matched pairs analysis ofthe relationship between generic strategy and performance. ServiceIndustries Journal, 13, 40–64.

Osterman, P. (1987). Choice of employment systems in internal labormarkets. Industrial Relations, 26, 46–67.

Ostroff, C., & Bowen, D. E. (2000). Moving human resource to a higherlevel: Human resource practices and organizational effectiveness. InK. J. Klein & S. W. J. Kozlowski (Eds.), Multilevel theory, research, andmethods in organizations (pp. 211–266). San Francisco: Jossey-Bass.

Parasuraman, A., Zeithaml, V. A., & Berry, L. L. (1988). SERVQUAL: Amultiple-item scale for measuring consumer perceptions of service qual-ity. Journal of Retailing, 64, 12–40.

Parasuraman, A., Zeithaml, V. A., & Berry, L. L. (1994). Reassessment ofexpectations as a comparison standard in measuring service quality:Implications for further research. Journal of Marketing, 58, 111.

Pfeffer, J. (1998). The human equation: Building profits by putting peoplefirst. Boston: Harvard Business School Press.

Pfeffer, J., & Davis-Blake, A. (1987). The effect of the proportion ofwomen on salaries: The case of college administrators. AdministrativeScience Quarterly, 32, 1–24.

Ployhart, R. E., Weekley, J. A., & Baughman, K. (2006). The structure andfunction of human capital emergence: A multilevel examination of theattraction–selection–attrition model. Academy of Management Journal,49, 661–677.

Porter, M. E. (1980). Competitive strategy. New York: Free Press.Pugh, S. D. (2001). Service with a smile: Emotional contagion in the

service encounter. Academy of Management Journal, 44, 1018–1027.Reichheld, F. (1996). The loyalty effect. Boston: Harvard Business School

Press.Rousseau, D. M. (1995). Psychological contracts in organizations. Thou-

sand Oaks, CA: Sage.Rousseau, D. M. (2005). I-deals: Idiosyncratic deals employees bargain

for themselves. Armonk, NY: Sharpe.Rust, R. T., Moorman, C., & Dickson, P. R. (2002). Getting return on

quality: Revenue expansion, cost reduction, or both? Journal of Mar-keting, 66, 7–24.

Ryan, R. M., Mims, V., & Koestner, R. (1983). Relation of rewardcontingency and interpersonal context to intrinsic motivation: A reviewand test using cognitive evaluation theory. Journal of Personality andSocial Psychology, 45, 736–750.

Schlesinger, L. A., & Heskitt, J. L. (1991). Breaking the cycle of failure inservices. Sloan Management Review, 17, 57–68.

Schneider, B. (1987). The people make the place. Personnel Psychology,40, 437–454.

Schneider, B. (1990). The climate for service: An application of the climateconstruct. In B. Schneider (Ed.), Organizational climate and culture:383–412. San Francisco: Jossey-Bass.

Schneider, B. (in press). The people still make the place. In D. B. Smith(Ed.), The people make the place: Dynamic linkages between people andorganizations—A festschrift in honor of Benjamin Schneider. Mahwah,NJ: Erlbaum.

Schneider, B., & Bowen, D. E. (1985). Employee and customer perceptions

of service in banks: Replication and extension. Journal of AppliedPsychology, 70, 423–433.

Schneider, B., Ehrhart, M. G., Mayer, D. M., Saltz, J. L., & Niles-Jolly, K.(2005). Understanding organization–customer links in service settings.Academy of Management Journal, 48, 1017–1032.

Schneider, B., Goldstein, H. W., & Smith, D. B. (1995). The ASA frame-work: An update. Personnel Psychology, 48, 747–773.

Schneider, B., White, S. S., & Paul, M. C. (1998). Linking service climateand customer perceptions of service quality: Tests of a causal model.Journal of Applied Psychology, 83, 150–163.

Seibert, S. E., Silver, S. R., & Randolph, W. A. (2004). Taking empow-erment to the next level: A multiple-level model of empowerment,performance, and satisfaction. Academy of Management Journal, 47,332–349.

Seiders, K., Voss, G. B., Grewal, D., & Godfrey, A. L. (2005). Do satisfiedcustomers buy more? Examining moderating influences in a retailingcontext. Journal of Marketing, 69, 26–43.

Settoon, R. P., Bennett, N., & Liden, R. C. (1996). Social exchange inorganizations: Perceived organizational support, leader–member ex-change, and employee reciprocity. Journal of Applied Psychology, 81,219–227.

Shore, L. M., & Tetrick, L. E. (1991). A construct validity study of thesurvey of perceived organizational support. Journal of Applied Psychol-ogy, 76, 637–643.

Shore, L. M., Tetrick, L. E., Taylor, M. S., Coyle-Shapiro, J., Liden, R.,McLean Parks, J., Morrison, E. W., Porter, L. W., Robinson, S., Roeh-ling, M., Rousseau, D., Schalk, R., Tsui, A., & Van Dyne, L. (2004). Theemployee–organization relationship: A timely concept in a period oftransition. In J. Martocchio (Ed.), Research in personnel and humanresources management (Vol. 23, pp. 291–370). Greenwich, CT: JAIPress.

Shore, L. M., & Wayne, S. J. (1993). Commitment and employee behavior:Comparison of affective commitment and continuance commitment withperceived organizational support. Journal of Applied Psychology, 78,774–780.

Shrout, P. E., & Bolger, N. (2002). Mediation in experimental and non-experimental studies: New procedures and recommendations. Psycho-logical Methods, 74, 422–445.

Snell, S. A., Youndt, M. A., & Wright, P. M. (1996). Establishing aframework for research in strategic human resource management: Merg-ing resource theory and organizational learning. In G. R. Ferris (Ed.),Research in personnel and human resources management (pp. 61–90).Greenwich, CT: JAI Press.

Sobel, M. E. (1982). Asymptotic confidence intervals for indirecteffects in structural equation models. In S. Leinhardt (Ed.), Socio-logical methodology (pp. 290 –312). Washington, DC: AmericanSociological Association.

Solomon, M. R., Surprenant, C., Czepiel, J. A., & Gutman, E. G. (1985).A role theory perspective on dyadic interactions: The service encounter.Journal of Marketing, 49, 99–111.

Spreitzer, G. M. (1995). Psychological empowerment in the workplace:Dimensions, measurement, and validation. Academy of ManagementJournal, 38, 1442–1465.

Subramaniam, M., & Youndt, M. A. (2005). The influence of intellectualcapital on the types of innovative capabilities. Academy of ManagementJournal, 48, 450–463.

Tetrick, L. E., Shore, L. M., Newton, L. A., & Vandenberg, R. J. (2007).A model of union participation: The impact of perceived union support,union instrumentality, and union loyalty. Journal of Applied Psychology,92, 820–828.

The world factbook. Washington, DC: Central Intelligence Agency. Re-trieved October 5, 2007 from https://www.cia.gov/library/publications/the-world-factbook/index.html

Thomas, K. W., & Velthouse, B. A. (1990). Cognitive elements of em-


powerment: An “interpretive” model of intrinsic task motivation. Acad-emy of Management Review, 15, 666–681.

Tyagi, P. K. (1985). Relative importance of key job dimensions andleadership behaviors in motivating salesperson work performance. Jour-nal of Marketing, 86, 76–86.

Wayne, S. J., Shore, L. M., Bommer, W. H., & Tetrick, L. E. (2002). Therole of fair treatment and rewards in perceptions of organizationalsupport and leader–member exchange. Journal of Applied Psychology,87, 590–598.

Wayne, S. J., Shore, L. M., & Liden, R. C. (1997). Perceived organiza-tional support and leader–member exchange: A social exchange per-spective. Academy of Management Journal, 40, 82–111.

Wright, P. M., & Boswell, W. R. (2002). Desegregating human resourcemanagement: A review and synthesis of micro and macro human re-source management research. Journal of Management, 24, 247–276.

Wright, P. M., & McMahan, G. C. (1992). Theoretical perspectives forstrategic human resource management. Journal of Management, 18,295–320.

Youndt, M. A., Subramaniam, O., & Snell, S. A. (2004). Intellectual capitalprofiles: An examination of investments and returns. Journal of Man-agement Studies, 41, 335–361.

Yuan, K. H., Bentler, P. M., & Kano, Y. (1997). On averaging variables ina confirmatory factor analysis model. Behaviormetrika, 24, 71–83.

Zacharatos, A., Barling, J., & Iverson, R. D. (2005). High-performancework systems and occupational safety. Journal of Applied Psychology,90, 77–93.

Received June 12, 2007Revision received June 16, 2008

Accepted June 27, 2008 �

E-Mail Notification of Your Latest Issue Online!

Would you like to know when the next issue of your favorite APA journal will be availableonline? This service is now available to you. Sign up at http://notify.apa.org/ and you will benotified by e-mail when issues of interest to you become available!


do they see eye to eye? management and employee ... · do they see eye to eye? management and...

Documents