Meta Analytic Analysis of SHRM

8/13/2019 Meta Analytic Analysis of SHRM

1/31

HOW DOES HUMAN RESOURCE MANAGEMENT INFLUENCEORGANIZATIONAL OUTCOMES?

A META-ANALYTIC INVESTIGATION OFMEDIATING MECHANISMS

KAIFENG JIANGDAVID P. LEPAK

Rutgers, the State University of New Jersey

JIA HUUniversity of Notre Dame

JUDITH C. BAERRutgers, the State University of New Jersey

Drawing on the ability-motivation-opportunity model, this meta-analysis examined theeffects of three dimensions of HR systemsskills-enhancing, motivation-enhancing,and opportunity-enhancingon proximal organizational outcomes (human capitaland motivation) and distal organizational outcomes (voluntary turnover, operationaloutcomes, and financial outcomes). The results indicate that skill-enhancing practiceswere more positively related to human capital and less positively related to employeemotivation than motivation-enhancing practices and opportunity-enhancing practices.Moreover, the three dimensions of HR systems were related to financial outcomes bothdirectly and indirectly by influencing human capital and employee motivation as wellas voluntary turnover and operational outcomes in sequence.

In the past two decades, researchers in strategichuman resource management (HRM) have exam-ined why and how organizations achieve theirgoals through the use of human resource (HR) prac-

tices. Although traditional HRM research has fo-cused on the impact of individual HR practices, thestrategic perspective on HRM research emphasizesbundles of HR practices, often referred to as high-performance work systems (HPWS), high-involve-ment work systems, and high-commitment worksystems, in examinations of the effects of HRM onemployee and organizational outcomes (Wright &McMahan, 1992). A burgeoning body of strategicHRM research has shown that the use of systemsof HR practices intended to enhance employeesknowledge, skills, and abilities, motivation, and

opportunity to contribute is associated with posi-tive outcomes such as greater commitment (Gong,Law, Chang, & Xin, 2009), lower turnover (Batt,

2002), higher productivity and quality (MacDuffie,1995), better service performance (Chuang & Liao,2010), enhanced safety performance (Zacharatos,Barling, & Iverson, 2005), and better financial per-formance (Huselid, 1995).

Despite the robust evidence for the positive rela-tionships between HRM and various organizationaloutcomes (Combs, Liu, Hall, & Ketchen, 2006), im-portant issues remain regarding the mechanismsthrough which HRM is associated with differentorganizational outcomes. First, the theoretical logicunderlying the mechanisms linking HRM andorganizational outcomes remains fragmented(Huselid & Becker, 2011; Wright & Gardner, 2003).Specifically, some researchers have adopted a be-havioral perspective to suggest that HR practicesaffect organizational outcomes by influencing em-ployee role behaviors; if employees act in ways thatare consistent with company goals, performanceshould improve. Other researchers have adoptedmore of a human capital and resource-based per-spective, focusing on the potential contributions ofemployees competenciesthat is, their knowl-edge, skills, and abilities. Interestingly, althoughemployees contribute through both their competen-cies and their actions, researchers have typicallyfocused on one perspective to understand how HR

We thank the action editor for this article, Jason Shaw,and three anonymous reviewers, Patrick McKay, RebeccaKehoe, and Mark Huselid for helpful comments and sug-gestions. We acknowledge financial support from theSHRM Foundation (Project No. 143). The interpretations,conclusions, and recommendations are those of the au-thors and do not necessarily represent those of the SHRMFoundation.

Academy of Management Journal

2012, Vol. 55, No. 6, 12641294.

http://dx.doi.org/10.5465/amj.2011.0088

1264Copyright of the Academy of Management, all rights reserved. Contents may not be copied, emailed, posted to a listserv, or otherwise transmitted without the copyright holders express

written permission. Users may print, download, or email articles for individual use only.


2/31

systems impact organizational outcomes (excep-tions include Takeuchi, Lepak, Wang, and Takeu-chi [2007]). Considering multiple perspectivessimultaneously provides a broader and more com-plete picture of the relationship between HRM andorganizational outcomes.

Second, although prior research has demonstrated

the mechanism through which HRM relates to someorganizational outcomes, it remains unclear as to howHRM relates to different organizational outcomes thatrange from very proximal (i.e., HR outcomes) to moredistal (i.e., financial outcomes). This lack of integra-tion is problematic given the different perspectivesadopted in the literature, perspectives that mighthighlight the importance of different but potentiallyrelated outcomes. Exploring the possible paths be-tween HRM and financial outcomes will likely pro-vide a more integrative model of how HR systemsoperate to impact a multitude of related and impor-tant outcomes (e.g., Becker & Huselid, 1998; Delery &Shaw, 2001; Guest, 1997).

Third, it is assumed in existing research that thecomponents of HR systems have identical impactson outcomes. For example, when scholars adopt anadditive approach to measure HR systems, eachcomponent of the system is treated as if it exerts anequal influence on the outcomes under investiga-tion. Although this is a possible reflection of howHR systems operate, scholars have recently chal-lenged this assumption and argued that differentsets of HR practices may impact the same outcomesin a heterogeneous way (e.g., Batt & Colvin, 2011;

Gardner, Wright, & Moynihan, 2011; Gong et al.,2009; Shaw, Dineen, Fang, & Vellella, 2009; Subra-mony, 2009). As these studies have suggested, it isimportant to explore the differential effects of thedifferent components of HR systems.

Given these issues, the primary objective of thisstudy is to develop an integrative model of the mech-anisms mediating between HRM and organizationaloutcomes through a meta-analytic approach. Drawingon the behavioral perspective on HRM, human capi-tal theory, and the resource-based view of the firm,we aim to extend and refine existing HRM-organiza-

tional outcomes models by exploring multiple medi-ating paths and differentiating among the effects ofsubdimensions of HR systems.

THEORETICAL BACKGROUND ANDHYPOTHESES

Existing Theories and Research on Relationshipsbetween HRM and Organizational Outcomes

Understanding the relationship between HRMand organizational outcomes is one of the long-

standing goals of macro HRM research. Indeed,Becker and Huselid (1998) considered this relation-ship as one of the essential pursuits of strategicHRM research. This stream of research has severalkey components. First, organizational outcomes areviewed as multidimensional. Drawing on Dyer andReevess (1995) work, researchers in strategic HRM

have categorized organizational outcomes intothree primary groups related to HRM: HR out-comes, operational outcomes, and financial out-comes. HR outcomes refer to those most directlyrelated to HRM in an organization, such as em-ployee skills and abilities, employee attitudes andbehaviors, and turnover.Operational outcomesarethose related to the goals of an organizational op-eration, including productivity, product quality,quality of service, and innovation. Financial out-comesreflect the fulfillment of the economic goalsof organizations. Typical financial outcomes in-clude sales growth, return on invested capital, andreturn on assets. In this study, we use organiza-tional outcomes to refer to all three categories ofoutcomes at the organizational level.

Second, strategic HRM research suggests that dif-ferent types of outcomes may not necessarily haveequivalent relationships with HR practices (Becker& Huselid, 1998; Delery & Shaw, 2001; Guest, 1997;Lepak, Liao, Chung, & Harden, 2006; Ostroff & Bo-wen, 2000). Moreover, it is commonly asserted thatHRM may influence the three types of organiza-tional outcomes in sequence. For example, HRpractices are expected to first influence HR out-

comes (e.g., employee skills and motivation),which are proximal and the least likely to be con-taminated by factors beyond HR practices. HR out-comes, in turn, may mediate the influence of HRpractices on productivity, quality, service, safety,innovation, and other operational outcomes, whichfurther affect financial outcomes.

Although existing HR research often implies thatHR outcomes serve as a key mediator between HRsystems and key outcomes, the specific natures ofmodels of this meditation depend on the theoreti-cal perspective researchers have adopted when ex-

amining this relationship. On the one hand, severalresearchers have adopted the behavioral perspec-tive of HRM (Jackson, Schuler, & Rivero, 1989).According to this perspective, organizations do notperform themselves, but instead use HR practicesto encourage productive behaviors from employeesand thus to achieve desirable operational and fi-nancial objectives (Becker & Huselid, 1998). If anorganization requires efficient employees, for ex-ample, its chosen HR practices and their effective-ness would likely differ from those of an organiza-tion that requires employees to be cooperative, to

2012 1265Jiang, Lepak, Hu, and Baer


3/31

focus on service, or to engage in some other criticalrole behavior. The effectiveness of HR practices isrealized when employees act in ways that areneeded for implementing strategies and achievingvarious business objectives.

On the other hand, some macro HRM researchershave focused less on the behaviors of employees

and more on their competencies within organiza-tions. Researchers taking on this perspective ofteninvoke human capital theory and the resource-based view of the firm. Human capital theory em-phasizes that human capitalthe composition ofemployee skills, knowledge, and abilitiesis a cen-tral driver of organizational performance when thereturn on investment in human capital exceeds la-bor costs (Becker, 1964; Lepak & Snell, 1999; Ploy-hart & Moliterno, 2011). The resource-based viewprovides additional insights as to why human cap-ital can help firms to outpace competitors and pro-poses that organizations obtain a competitive ad-vantage from resources that are rare, valuable,inimitable, and nonsubstitutable (Barney, 1991;Mahoney & Pandian, 1992). Researchers have ar-gued that human capital, especially high-qualityand/or organization-specific human capital, has thepotential to serve as a source of competitive advan-tage (Wright, McMahan, & McWilliams, 1994). Or-ganizations may use HR practices to create andmaintain valuable human capital, including bothgeneric and organization-specific human capital,which in turns drives high operational and finan-cial performance (Becker & Huselid, 1998; Delery &

Shaw, 2001; Ployhart & Moliterno, 2011; Snell &Dean, 1992).

Although the behavioral perspective of HRM, hu-man capital theory, and the resource-based view ofthe firm let researchers adopt different angles tolook at the relationships between HR practices andmore distal outcomes, under all three perspectivesHR outcomes are viewed as a critical path fromHRM to operational and financial outcomes. Evenwith this agreement, however, researchers have notsuccessfully combined multiple approaches to de-lineate an overarching picture of how this path

unfolds. For example, most of the extant empiricalresearch has examined the influence of HR systemson operational or financial performance eitherthrough motivation-related variables (e.g., Chuang& Liao, 2010; Collins & Smith, 2006; Gelade & Ivery,2003; Gong et al., 2009; McClean & Collins, 2011;Sun, Aryee, & Law, 2007) or through human capitalvariables (e.g., Cabello-Medina, Lopez-Cabrales, &Valle-Cabrera, 2011; Yang & Lin, 2009; Youndt &Snell, 2004). Insights into each type of variable areimportant yet insufficient to fully capture the pro-cess linking HRM to outcomes. Thus, research is

needed to explore how HRM can help organiza-tions achieve financial goals through multiplepaths (Takeuchi et al., 2007).

Decomposing HR Systems intoThree HR Dimensions

Scholars have recently argued that although em-ployees are exposed to HR systems rather than in-dividual practices, the parts of these systems arenot necessarily equivalent in their impact. Mostresearch has portrayed an HR system as an additiveindex of a set of individual HR practices (Combs etal., 2006); there are reasons to believe, however,that the highly varied set of HR practices can becategorized into several subdimensions. Indeed,dividing HR systems into subdimensions is notnew in strategic HRM research. For example, draw-ing on an employee-organization relationshipframework (Tsui, Pearce, Porter, & Tripoli, 1997),researchers have argued that HR practices may becategorized as falling into HRM inducement andinvestment practices, and HRM expectation-enhancing practices (e.g., Batt & Colvin, 2011; Gonget al., 2009; Shaw et al., 2009; Shaw, Delery, Jen-kins, & Gupta, 1998; Shaw, Gupta, & Delery, 2005).The first two types are designed to improve em-ployees expected outcomes, whereas the third typereflects organizations expectations about employ-ees contributions.

Taking a different approach, some researchershave drawn upon the ability-motivation-opportu-

nity (AMO) model of HRM and suggested that em-ployee performance is a function of three essentialcomponents: ability, motivation, and opportunityto perform. Extending this logic, HR systems de-signed to maximize employee performance can beviewed as a composition of three dimensions in-tended to enhance employee skills, motivation, andopportunity to contribute, respectively (Appel-baum, Bailey, Berg, & Kalleberg, 2000; Bailey, 1993;Boxall & Purcell, 2008; Delery & Shaw, 2001; Ger-hart, 2007; Katz, Kochan, & Weber, 1985; Lepak etal., 2006). Recently, several empirical studies have

adopted and validated this conceptual framework(e.g., Bailey, Berg, & Sandy, 2001; Batt, 2002; Gard-ner et al., 2011; Huselid, 1995; Kehoe & Wright, inpress; Liao, Toya, Lepak, & Hong, 2009; MacDuffie,1995; Subramony, 2009).

In keeping with these studies, Lepak and col-leagues (2006) suggested that it might be fruitful toconceptualize HR practices as falling into one ofthree primary dimensions: skill-enhancing HRpractices, motivation-enhancing HR practices,and opportunity-enhancing HR practices. Skill-enhancing HR practicesare designed to ensure ap-

1266 DecemberAcademy of Management Journal


4/31

propriately skilled employees; they include com-prehensive recruitment, rigorous selection, andextensive training.Motivation-enhancing HR prac-ticesare implemented to enhance employee moti-vation. Typical ones include developmental perfor-mance management, competitive compensation,incentives and rewards, extensive benefits, promo-

tion and career development, and job security.Op-portunity-enhancing HR practicesare designed toempower employees to use their skills and motiva-tion to achieve organizational objectives. Practicessuch as flexible job design, work teams, employeeinvolvement, and information sharing are generallyused to offer these opportunities. The use of thethree dimensions of HR systems instead of a unidi-mensional or two-dimensional framework is basedon an examination of differential effects of the threedimensions of HR systems on different types of HRoutcomes.

Linking HR Dimensions to Multiple Outcomes

According to the ability-motivation-opportunitymodel of HRM, HR outcomes can conceptually bedivided into human capital, motivation, and oppor-tunity to contribute (Becker & Huselid, 1998; Del-ery & Shaw, 2001; Guest, 1997), and human capitaland employee motivation are two of the most crit-ical mediating factors that have been examined inthe literature (e.g., Gardner et al., 2011; Gong et al.,2009; Liao et al., 2009; Sun et al., 2007; Takeuchi etal., 2007; Youndt & Snell, 2004). In line with the

literature, we focus on the mediating roles of hu-man capital and employee motivation. As previousresearch suggests, human capital can be viewed asa composition of employees knowledge, skills, andabilities (Coff, 2002), and employee motivation re-fers to the direction, intensity, and duration of em-ployees effort (Campbell, McCloy, Oppler, & Sager,1993), as manifested by positive work attitudes(e.g., collective job satisfaction, commitment, per-ceived organizational support) and work behaviors(e.g., organizational citizenship behavior).

Although we anticipate that all three HR dimen-

sions are positively related to both human capitaland employee motivation, we also anticipate thatthe three HR dimensions may play different roles inbuilding human capital and enhancing employeemotivation. We expect that, compared with moti-vation-enhancing and opportunity-enhancing HRpractices, skill-enhancing HR practices will likelyhave a stronger impact on human capital and aweaker impact on employee motivation.

According to the ability-motivation-opportunityframework, skill-enhancing HR practices can di-rectly help to optimize the levels or types of em-

ployees skills and abilities. For example, recruit-ment and selection practices are intended to insurethat employees have the skills needed for task per-formance, and training and development may fur-ther provide employees with organization-specificskills with which to perform their work. Indeed,Delaney and Huselid (1996) indicated that organi-

zations can enhance the skills of their workforcesboth by hiring high-quality individuals and by im-proving the level of skills in their current work-forces. Relatedly, prior research shows that the useof comprehensive selection and training practicesfostered employees collective human capital (e.g.,Cabello-Medina et al., 2011; Takeuchi et al., 2007;Yang & Lin, 2009; Youndt & Snell, 2004). Further-more, research suggests that practices such as com-petitive compensation, extensive benefits, and jobsecurity may help attract capable employees andretain them in organizations, and practices such aswork teams, employee involvement, and flexiblejob design may provide employees with opportuni-ties to share knowledge and to learn new skills.However, the relationships between the other twoHR dimensions and human capital are seen as lessdirect. Research has shown that practices fromthese two dimensions were less positively relatedto human capital than skill-enhancing HR practices(Cabello-Medina et al., 2011; Yang & Lin, 2009).Therefore, we propose the following:

Hypothesis 1a. Skill-enhancing HR practicesare positively related to human capital.

Hypothesis 1b. Motivation-enhancing HR prac-tices are positively related to human capital.

Hypothesis 1c. Opportunity-enhancing HRpractices are positively related to humancapital.

Hypothesis 2a. Skill-enhancing HR practicesare more positively related to human capitalthan motivation-enhancing HR practices.

Hypothesis 2b. Skill-enhancing HR practicesare more positively related to human capitalthan opportunity-enhancing HR practices.

We also posit that the three dimensions of HRsystems are positively related to employee motiva-tion to different degrees. First, investment in allthree HR dimensions generally indicates that organ-izations value and support employees contribu-tions. According to social exchange theory (Blau,1964) and the norm of reciprocity (Gouldner, 1960),employees who perceive an organizations actionstoward them as beneficial may feel obligated toreciprocate and be motivated to exert more effort atwork. More specifically, motivation-enhancing HR



5/31

practices (e.g., performance-based compensation,incentives and benefit, promotion opportunities,and job security) are more likely to provide em-ployees with extrinsic motivation that links theirwork efforts to external rewards. Practices such aswork teams, employee involvement, and flexiblejob design help to generate employees intrinsic

motivation, which encourages them to seek outchallenges at work (Ryan & Deci, 2000). In addition,skill-enhancing HR practices can enhance employ-ees skills and abilities, which may help careerdevelopment and induce promotion opportunitiesin their organizations (Tharenou, Saks, & Moore,2007). However, the effect of skill-enhancing HRpractices on employee motivation is relatively in-direct and likely to be contingent on the practicesin the other two HR dimensions. For example, eventhough training can improve employees skills atwork, the increased skills may not necessarily leadto promotion in their organization. Therefore, weexpect all three HR dimensions to be positivelyassociated with employee motivation and, com-pared with the other two dimensions, skill-enhanc-ing HR practices are less positively related to em-ployee motivation. Recent empirical research thatexamined the influence of three HR dimensions onemployee affective commitment (Gardner et al.,2011) has also supported this reasoning. Therefore,we hypothesize:

Hypothesis 3a. Skill-enhancing HR practicesare positively related to employee motivation.

Hypothesis 3b. Motivation-enhancing HR prac-tices are positively related to employeemotivation.

Hypothesis 3c. Opportunity-enhancing HRpractices are positively related to employeemotivation.

Hypothesis 4a. Skill-enhancing HR practicesare less positively related to employee motiva-tion than motivation-enhancing HR practices.

Hypothesis 4b. Skill-enhancing HR practicesare less positively related to employee motiva-

tion than opportunity-enhancing HR practices.

In addition to the direct effects of the three HRdimensions on human capital and employee moti-vation, we propose that human capital and em-ployee motivation mediate the relationships be-tween the three HR dimensions and more distaloutcomes related to voluntary turnover (voluntaryorganizational exit), operational outcomes, andsubsequent financial outcomes.

Several researchers have viewed voluntary turn-over as a critical intermediate outcome that is dis-

tinct from human capital and employee motivation(e.g., Batt, 2002; Batt & Colvin, 2011; Gardner et al.,2011; Guthrie, 2001; Shaw et al., 1998, 2005, 2009;Sun et al., 2007). Research has consistently demon-strated that HR practices designed to enhance em-ployee skills and motivation are significantly andnegatively associated with voluntary turnover (e.g.,

Arthur, 1994; Batt, 2002; Guthrie, 2001; Huselid,1995). Some researchers attribute the negative rela-tionships to the emotional bond between employ-ees and organizations formed by HR practices. Inother words, because HR practices enhance em-ployees motivation at work, these employees arereluctant to leave their organizations (e.g., Gardneret al., 2011; Sun et al., 2007). Investment in thethree aspects of HR systems implies that organiza-tions value employees contribution and expect toestablish long-term employment relationships withtheir employees. As a result, employees are encour-aged to work harder to reciprocate and thus are lessprone to quit their jobs.

Human capital theory and the resource-basedview of the firm indicate that employees with ap-propriate human capital resulting from HR invest-ments may be less likely to leave their organiza-tions. First, researchers have suggested thatemployees with high levels of human capital aremore capable of meeting job demands, receivingpositive performance appraisals, obtaining promo-tions, and participating in decision making (Batt &Colvin, 2011; Shaw et al., 2009). Therefore, com-pared with those with less human capital, employ-

ees with higher levels of human capital will be lesslikely to leave their organizations. In addition, em-ployees with high levels of human capital are betterable to learn at work, which facilitates the devel-opment of specific human capital (Ployhart & Mo-literno, 2011). The accumulated specific humancapital may in turn reduce the likelihood employ-ees leave, because the specific human capital that isunique and valuable for their current organizationmay not provide value to other organizations (Bar-ney, 1991; Lepak & Snell, 1999). Employees areunable to obtain return on their input in developing

the specific human capital if they quit (Shaw et al.,2005). Therefore, we hypothesize:

Hypothesis 5a. Human capital mediates thenegative relationships between the three di-mensions of HR systems and voluntaryturnover.

Hypothesis 5b. Employee motivation mediatesthe negative relationships between the threedimensions of HR systems and voluntaryturnover.



6/31

Human capital and employee motivation are alsoexpected to mediate the influence of the three HRdimensions on operational outcomes. Researchershave widely recognized the potential impact of hu-man capital on organizational effectiveness (Bar-ney, 1991; Coff, 1997; Snell, Youndt, & Wright,1996; Wright et al., 1994; Wright, Dunford, & Snell,

2001). According to human capital theory and theresource-based view, human capital is the primarydeterminant of productivity (Dess & Shaw, 2001)and can be a source of competitive advantage whenit is valuable and unique for an organization, hardto replace without significant costs, and not easilyimitated by rivals (Coff, 1997; Wright et al., 1994).Therefore, with high-quality human capital pools,organizations are more likely to achieve opera-tional goals such as high productivity and quality,great service, and innovation. Research has pro-vided support for the positive effect of human cap-ital on operational performance (Crook, Todd,Combs, Woehr, & Ketchen, 2011).

Moreover, researchers taking a behavioral per-spective suggest that the value of employees hu-man capital cannot be realized unless they are will-ing to use their capabilities (Jackson & Schuler,1995). To encourage employees to do so, organiza-tions need to utilize HR practices to enhance theirintrinsic and extrinsic motivation at work, whichcan further lead to desired work behaviors anddiscretionary efforts contributing to operationaloutcomes (Deci, Connell, & Ryan, 1989). A numberof empirical studies have shown that positive work

attitudes (e.g., collective commitment) and positiveperceptions of a work environment (e.g., perceivedorganizational support) mediate the relationshipsbetween high-performance work systems and oper-ational outcomes (e.g., Chuang & Liao, 2010; Gelade& Ivery, 2003; Rogg, Schmidt, Shull, & Schmitt,2001; Sun et al., 2007). Therefore, we hypothesize:

Hypothesis 6a. Human capital mediates thepositive relationships between the three di-mensions of HR systems and operationaloutcomes.

Hypothesis 6b. Employee motivation mediatesthe positive relationships between the three di-mensions of HR systems and operationaloutcomes.

Finally, we propose mediating effects of volun-tary turnover and operational outcomes on the re-lationships between the three HR dimensions andfinancial outcomes. The relationship between vol-untary turnover and financial performance is com-plex, depending on what kinds of employees leaveand whether they have been replaced appropri-

ately. According to human capital theory, whencapable employees leave, an organization loses thehuman capital embodied in those departing andalso loses the chance to realize a return on itsinvestment in developing the human capital (Dess& Shaw, 2001). Especially when employees possessorganization-specific human capital, the loss will

be detrimental for organizations financial perfor-mance, and organizations need to take a long timeto regain their competitive advantage (Osterman,1987; Strober, 1990). On the other hand, researchhas also suggested that organizations need somelevel of voluntary turnover. This is because em-ployees who do not fit their jobs will self-select outof organizations, which also need new employeesto provide fresh stimulus (Dalton & Todor, 1979;Jovanovic, 1979; Schneider, 1978). However, nomatter which kinds of employees leave, organiza-tions also incur additional costs related to turnover(Dess & Shaw, 2001). For example, administrativeresources used in recruitment, selection, and train-ing would have been in vain, and the organizationsneed to invest additional resources to search forand train new employees to replace the leavers. Atthe same time, operational outcomes will sufferduring the vacant and training period. Further, ahigh turnover rate can corrupt the morale of organ-izations and trigger more employees to leave theirjobs, thereby negatively affecting financial out-comes (Hausknecht & Trevor, 2011). In keepingthese arguments, empirical studies have consis-tently demonstrated the existence of a negative re-

lationship between voluntary turnover and finan-cial performance (e.g., Batt, 2002; Glebbeek & Bax,2004; Huselid, 1995; Kacmar, Andrews, Van Rooy,Steilberg, & Cerrone, 2006; Morrow & McElroy,2007; Shaw et al., 2005). Therefore, we propose anegative relationship between voluntary turnoverand financial performance.

The rationale for the positive relationship be-tween operational outcomes and financial out-comes is clear in the literature. The financial out-comes of an organization are a function of a varietyof factors, including industry environment, organ-

izational strategy, and organizational characteris-tics (White & Hamermesh, 1981). Among these ex-planatory factors, business operations within anorganization may be a salient determinant of finan-cial outcomes because outcomes such as productiv-ity, quality, and service are directly related to prof-itability (Curtis, Hefley, & Miller, 1995). In a meta-analytic review, Capon, Farley, and Hoenig (1990)found that quality of product and service were pos-itively associated with financial outcomes. Like-wise, Crook and colleagues (2011) also reported apositive relationship between operational out-



7/31

comes and financial outcomes. In view of thesefindings, we propose a positive relationship be-tween operational and financial outcomes.

In sum, drawing upon the behavioral perspectiveof HRM, human capital theory, and the resource-based view of the firm, we propose a mediatingmodel in which the three dimensions of HR sys-

tems are indirectly related to financial outcomesthrough human capital, employee motivation, vol-untary turnover, and operational outcomes in se-quence. In building this framework, we focus onthe mediating role of employees in the link of HRMwith financial performance. However, our modeldoes not exclude other paths through which HRMcan help increase financial outcomes. In fact, boththeoretical and empirical research has suggestedthat HRM can provide firms with organizationalcapital reflected by internal fit and flexibility(Evans & Davis, 2005; Wright & Snell, 1998) andsocial capital (Collins & Clark, 2003; Delery &Shaw, 2001; Gittell, Seidner, & Wimbush, 2010),both of which can be sources of competitive advan-tage for organizations. Given these alternative pos-sibilities, we hypothesize that the intermediate out-comes proposed in our model partially mediate thepositive relationships between the three HR dimen-sions and financial outcomes.

Hypothesis 7. Human capital, employee moti-vation, voluntary turnover, and operationaloutcomes partially mediate the positive rela-tionships between the three dimensions of HR

systems and financial outcomes.

METHODS

Data Collection

We tested the mediating hypotheses with thehelp of meta-analytic structural equation modeling(SEM) techniques (Cheung & Chan, 2005, 2009;Viswesvaran & Ones, 1995). To identify studies thatcould be used in the meta-analysis, we firstsearched the PsycINFO, Web of Science, and Pro-Quest Digital Dissertations databases for studies

published before May 2011. We used multiple key-words. For HRM, we used the keywords humanresource work practice/system, high-perfor-mance work practice/system, high-involvementwork practice/system, or high-commitment workpractice/system, whereas for organizational out-comes, we searched for studies that also includedthe keywords performance, outcome, atti-tudes, satisfaction, commitment, motiva-tion, human capital, turnover, productivity,quality, service, safety, growth, or profit-ability. Moreover, we used the same search terms

to search conference programs from the Academyof Management (AOM) and the Society of Indus-trial and Organizational Psychology from 2000 to2010. Second, we referred to the reference lists ofthe prior reviews on this topic, including theoreti-cal reviews (e.g., Becker & Gerhart, 1996; Becker &Huselid, 1998; Lengnick-Hall, Lengnick-Hall, An-

drade, & Drake, 2009; Lepak et al., 2006; Wright &Boswell, 2002) and meta-analytic reviews (Combset al., 2006; Subramony, 2009). Third, we made aneffort to identify unpublished studies through thelistservs of the AOMs Human Resources and Organ-izational Behavior Divisions.

Four inclusion criteria were used to select stud-ies. First, we focused only on studies that examinedthe relationships between HR practices and organ-izational outcomes at the organizational level (e.g.,establishment, business unit, or firm). We did notinclude studies that investigated individual-levelrelationships between employee-perceived HRpractices/systems and individual outcomes (e.g.,Agarwala, 2003; Barling, Kelloway, & Iverson,2003) or cross-level relationships between organi-zation-level HR practices and individual-level out-comes (e.g., Liao et al., 2009; Takeuchi, Chen, &Lepak, 2009). Second, we only included studiesthat emphasized the use of HR practices/systems inorganizations but not the effectiveness or the valueof these practices or systems (e.g., Huselid, Jackson,& Schuler, 1997; Richard & Johnson, 2004). Third,we included studies in the meta-analysis if theyreported at least one correlation among individual

HR practices and various organizational outcomes.We excluded the studies that only presented thecorrelations of HR systems rather than those ofindividual HR practices with organizational out-comes (e.g., Bae & Lawler, 2000). Studies withoutthe statistical information (e.g., sample sizes, cor-relation coefficients) necessary to calculate effectsizes were also excluded (e.g., Cappelli & Neumark,2001; Ichniowski, Shaw, & Prennushi, 1997). Fi-nally, when the same sample was used in two ormore articles, we considered only the one that pro-vided more information. In contrast, when a study

used two or more independent samples, we codedthese independent samples separately. The inclu-sion criteria yielded a final set of 116 articles rep-resenting 120 independent samples that included atotal of 31,463 organizations.

We first developed the coding sheet and instruc-tions as recommended by Lipsey and Wilson(2001). The first author and the third author thenindependently coded a random selection of 15 ar-ticles to assess the level of agreement regardingsample sizes, effect sizes, and reliability. After bothcoders checked data entry and resolved errors, they



8/31

independently coded the rest of studies. The con-sensus rate was 96 percent, and disagreementswere solved through discussion between the twocoders.

Operationalization of Variables

Three dimensions of HR systems.We identified14 HR practices frequently examined in the litera-ture. By following previous research using the abil-ity-motivation-opportunity framework (e.g., Appel-baum et al., 2000; Batt, 2002; Gardner et al., 2011;Guest, 1997; Lepak et al., 2006; Subramony, 2009),we categorized these practices into three dimen-sions. Skill-enhancing HR practices included re-cruitment, selection, and training. Motivation-en-hancing HR practices consisted of performanceappraisal, compensation, incentive, benefit, pro-motion and career development, and job security.In addition, opportunity-enhancing HR practicescovered job design, work teams, employee involve-ment, formal grievance and complaint processes,and information sharing.

Organizational outcomes.We summarized var-ious organizational outcomes into five categories.Human capital included overall organizational hu-man capital measured via established scales (e.g.,Subramaniam & Youndt, 2005; Youndt, Subrama-niam, & Snell, 2004) and the education level of aworkforce. Employee motivation was reflected bycollective job satisfaction, organizational commit-ment, organizational climate, perceived organiza-

tional support, and organizational citizenship be-havior. Voluntary turnover only represented thepercentage of employees who quit or voluntarilyleft the organizations. Dismissal rate and overallturnover rate were not included. In addition, weviewed productivity, quality, service, innovation,and overall operational performance as operationaloutcomes, and we viewed return on assets, returnon equity, market return, sale growth, and overallfinancial performance as financial outcomes.

As suggested by Aguinis, Pierce, Bosco, Dalton,and Dalton (2011), we provide a table, in Appendix

A, that lists all the included studies and our cate-gorizations of the three HR dimensions and differ-ent types of outcomes. This information is impor-tant to allow future research to replicate and extendthis study.

Meta-analytic and Model-Testing Procedures

To test the mediating model through meta-ana-lytic SEM, we needed to calculate meta-analyticcorrelations among three dimensions of HR sys-tems and different types of organizational out-

comes by correcting for measurement error andsampling error (Hunter & Schmidt, 2004). We firstperformed reliability corrections for informant-re-ported measures of HR practices and organizationaloutcomes to correct for measurement error. Forthose studies that did not report the reliabilities ofthe informant-reported measures, we imputed the

reliabilities using the weighted mean of the avail-able reliabilities estimated from the other studies(Lipsey & Wilson, 2001). Regarding the variablesthat were measured with archival data (e.g., returnon assets), we adopted a more conservative .80reliability estimate, which has been used in previ-ous meta-analyses in management (e.g., Dalton,Daily, Certo, & Roengpitya, 2003; Dalton, Daily,Ellstrand, & Johnson, 1998; Dalton, Daily, Johnson,& Ellstrand, 1999). For example, if training prac-tices were measured by reflective items (e.g., Thisfirm invests considerable time and money in train-

ing) in a study that reported the reliability of theseitems, we would correct for the reliability for train-ing. In contrast, if training practices were measuredby archival data (e.g., On average how many hoursof formal training do employees in this firm receiveeach year?), we would correct for a reliability of.80 for this measure. For comparison purposes, wealso calculated the reliability-corrected correlationby using a reliability of 1.00 for archival measuresand did not find changes in the main findings ofthis study.

Second, to calculate the composites of HR prac-tices (i.e., HR dimensions) and the composites ofoutcome variables (i.e., organizational outcomescategories), we combined the correlations amongindividual HR practices and outcomes using theformula provided by Hunter and Schmidt (2004:435439):

rXY rxiyj

nnn1 rxixjmmm1 ryiyj.

If it is assumed thatxrepresents a dimension of HRsystems (e.g., skill-enhancing HR practices) and yrepresents a category of organizational outcomes(e.g., employee motivation), rxiyjis the sum of the

correlations between HR practices (e.g., recruit-ment, selection, and training) and outcome vari-ables (e.g., collective satisfaction and commit-ment); nand mare the numbers of HR practicesand outcome variables respectively;rxixjis the av-erage correlation among HR practices; and ryiyjisthe average correlation among outcome variables.By using this formula, we created a single effectsize for each relationship within each study.

Third, we used a random-effects model to correctfor the sampling error by weighting each studyseffect size by its sample size (Hunter & Schmidt,



9/31

2004). We also computed the 95% confidence in-terval (CI) around the sample-weighted mean cor-relation and Qhomogeneity statistic. Confidenceintervals provide an estimate of the variabilityaround the estimated average correlation; a 95% CIexcluding zero indicates that one can be 95 percentconfident that the confidence interval includes the

average mean true score. The Qstatistic indicatesthe variance in the sample-weighted mean correla-tion; a significantQsuggests the heterogeneity of agiven relationship. Research has suggested that arandom-effects model provides a more accurate es-timate than a fixed-effects model when relation-ships are heterogeneous (Cheung & Chan, 2005;Erez, Bloom, & Wells, 1996; Overton, 1998).

Finally, we used the created correlation matricesin SEM computed in LISREL 8.72 (Jreskog & Sr-bom, 2005). Because the sample sizes for differentcorrelations were not identical, we imputed thesample size for the SEM analyses by calculating theharmonic mean of the correlation sample sizes(Viswesaran & Ones, 1995). Compared with thearithmetic mean, the harmonic mean gives muchless weight to large sample sizes and thus results ina more conservative parameter estimate. Four es-tablished model fit statisticschi-square (2), theroot-mean-square error of approximation (RMSEA),the comparative fit index (CFI), and the standard-ized root-mean-square residual (SRMR)wereused to examine the viability of the structural mod-els (Kline, 2005). Acceptable model fit is associatedwith nonsignificant chi-square values and with a

CFI greater than .90, an RMSEA less than or equalto .08, and an SRMR less than .10 (Kline, 2005). Weused two statistics to test the hypotheses predictingrelative effects of three HR dimensions on humancapital and employee motivation. One was the Z-test, which shows the significance of the differencebetween regression coefficients (Clogg, Petkova, &Haritou, 1995), and the other was the epsilon sta-tistic, which has been commonly used to determinethe relative weight of each predictor in explainingthe variance of dependent variables (Johnson, 2000;Johnson & LeBreton, 2004). The results of relative

weights represent the proportion of total variance(R2) explained by each HR dimension. To analyzemediation, we used Sobels (1982) test to examinethe statistical significance of indirect effects.

RESULTS

Differential Effects of HR Dimensions

Table 1 summarizes the correlation results of therelationships among HR dimensions and organiza-tional outcomes categories. To test Hypotheses 1, 2,

3, and 4, we included all three dimensions of HRsystems in regressions examining their effects onhuman capital and employee motivation. As shownin Table 2, all three HR dimensions had significantand positive effects on human capital. The resultsof Z-tests show that the regression coefficient ofskill-enhancing HR practices ( .29,p .01) was

significantly larger than the coefficients of motiva-tion-enhancing HR practices ( .22,p .01,Z 2.74,p .01) and opportunity-enhancing HR prac-tices ( .07,p .01,Z 8.68,p .01). Moreover,the analyses of relative weights indicate that skill-enhancing HR practices explained the largest per-centage of variance in human capital (48%), fol-lowed by motivation-enhancing HR practices(36%) and opportunity-enhancing HR prac-tices (16%).

Similarly, we found significantly positive effectsof three HR dimensions on employee motivation.Consistently with our prediction, the influences ofmotivation-enhancing HR practices ( .29, p .01,Z 8.64,p .01) and opportunity-enhanc-ing HR practices ( .25, p .01,Z 7.07,p .01) were significantly stronger than that of skill-enhancing HR practices ( .07,p .01). Motiva-tion-enhancing HR practices and opportunity-en-hancing HR practices respectively explained 45and 38 percent of the variance of employee moti-vation, whereas skill-enhancing HR practices ex-plained 17 percent. In sum, Hypotheses 1 through 4were supported.

Mediation Results

Hypotheses 5 through 7 predict that the three HRdimensions have both direct effects and indirecteffects through human capital, employee motiva-tion, voluntary turnover, and operational outcomeson financial outcomes. We tested the proposedmodel (Figure 1) by inputting correlation matrices(Table 1) into LISREL 8.72 (Jreskog & Srbom,2005). As shown in Table 3, the model fit of theproposed model was acceptable (2[9] 264.82,RMSEA .09, CFI .98, SRMR .04). All the

proposed relationships among HR dimensions andorganizational outcomes categories were signifi-cant and consistent with our prediction exceptfor the direct relationship between opportunity-enhancing HR practices and financial outcomes( .03, n.s.). Thus, we dropped this direct pathfrom the model, which only marginally impactedfit (model 1: 2[1] 4.65,p .05). We also testedthe direct relationships between three HR dimen-sions and voluntary turnover and operational out-comes. As presented in Table 3, adding paths fromskill-enhancing HR practices to both outcomes sig-



10/31

nificantly improved fit over that of model 1 (model2: 2[2] 80.71,p .01). However, the path fromskill-enhancing HR practices to voluntary turnoverwas not significant ( .02, p .05). Droppingthis path did not impact fit (model 3: 2[1] 1.56,

n.s.). Furthermore, we added the direct paths frommotivation-enhancing HR practices to voluntaryturnover and operational outcomes and found asignificant improvement in the fit over that ofmodel 3 (model 4: 2[2] 10.54, p .01). The

TABLE 1Meta-analytic Correlations between HR Dimensions and Organizational Outcomesa

Variables 1 2 3 4 5 6 7

1. Skill-enhancing practices2. Motivation-enhancing practices (r, r

c) .38, 46

k(N) 55 (14,670)95% CI .40: .53

Q 822.75**3. Opportunity-enhancing practices (r, r

c) .38, 47 .37, 44

k(N) 49 (13,079) 50 (13,740)95% CI .40: .53 .37: .52Q 557.95** 855.49**

4. Human capital (r, rc) .35, 42 .36, 38 .25, 30

k(N) 13 (2,013) 19 (3,249) 13 (2,068)95% CI .27: .57 .26: .49 .24: .37Q 133.88** 175.97** 23.58*

5. Employee motivation (r, rc) .25, 32 .33, 43 .32, 41 .37, 42

k(N) 20 (4,915) 22 (4,591) 19 (4,647) 12 (1,165)95% CI .26: .37 .34: .51 .31: .51 .23: .61Q 63.23** 148.79** 183.61** 111.91**

6. Voluntary turnover (r, rc) .19, .21 .15, .17 .17, .22 .22, .26 .31, .37

k(N) 19 (6,181) 24 (6,674) 19 (8,092) 7 (1,363) 11 (2,879)95% CI .14: .29 .09: .25 .10: .33 .01: .53 .18: .56Q 142.39** 213.38** 384.16** 130.46** 208.62**

7. Operational outcomes (r,rc) .25, 32 .19, 25 .25, 32 .25, 29 .32, 38 .15, .19

k(N) 36 (10,224) 37 (11,041) 35 (9,576) 8 (1,198) 23 (4,618) 22 (6,002)95% CI .25: .39 .17: .33 .25: .39 .06: .53 .30: .47 .10: .27Q 436.07** 626.61** 354.35** 108.92** 142.24** 189.14**

8. Financial outcomes (r, rc) .22, 26 .22, 27 .15, 20 .19, 24 .32, 38 .15, .19 .38, 48

k(N) 41 (9,966) 41 (12,219) 27 (5,610) 12 (2,028) 17 (3,354) 17 (4,055) 33 (8,863)95% CI .21: .32 .21: .33 .13: .26 .16: .32 .25: .51 .08: .30 .39: .57Q 247.07** 384.70** 130.73** 32.04** 206.61** 181.60** 547.24**

a The mean sample-size-weighted correlation (r) and mean sample-sized-weighted correlation corrected for attenuation due to unreli-ability (r

c) are presented. A k indicates the number of independent samples, and N is the total sample size. The 95% CI is the

95% confidence interval around the mean sample-size-weighted corrected correlation (rc).Qis the chi-square-test for the homogeneity of

corrected correlations (rc

) across studies.*p .05**p .01

TABLE 2Results of Differential Effects of HR Dimensions on Human Capital and Motivation-Related Attitudesa

Predictors

Human Capital Employee Motivation

t %R2 t %R2

Skill-enhancing HR practices (A) .29 15.76** 48% .07 3.90** 17%Motivation-enhancing HR practices (M) .22 12.12** 36% .29 16.30** 45%Opportunity-enhancing HR practices (O) .07 3.84** 16% .25 14.12** 38%Total R2 .22 .25Z, AM 2.74** 8.64**Z, AO 8.68** 7.07**

a Standardized coefficients are presented. Zis the test for the significance of the difference between the regression coefficients.*p .05

**p .01



11/31

FIGURE 1Theoretical Model of Effects of HR Dimensions on Organizational Outcomes

Skill-Enhancing

HR Practices

Motivation-

Enhancing HR

Practices

Opportunity-

Enhancing HR

Practices

Human

Capital

Employee

Motivation

Voluntary

Turnover

OperationalOutcomes

Financial

Outcomes

TABLE 3

Fit Statistics for Alternative Modelsa

Models 2 df 2 CFI RMSEA SRMR AIC

Three HR dimensionsTheoretical model (Figure 1) 264.82 9 .98 .09 .04 318.32Alternative model 1b 269.47 10 4.65*c .98 .08 .05 321.47Alternative model 2d 188.76 8 80.71**e .98 .08 .03 244.76Alternative model 3f 190.32 9 1.56e .98 .07 .03 244.32Alternative model 4g 179.78 7 10.54**e .99 .08 .03 237.78Alternative model 5h 180.54 8 0.76e .99 .08 .03 236.54Alternative model 6i (Figure 2) 130.32 6 50.22**e .99 .08 .02 190.32

Latent high performance work systems (HPWS)Theoretical model (Figure 3) 570.74 16 .95 .10 .05 610.74Alternative model 7j (Figure 4) 406.51 14 147.72**k .96 .09 .03 450.51

a n 3,724.b Deletes the direct paths from opportunity-enhancing HR practices to financial outcomes.c Model fit compared with the theoretical model of the effects of three HR dimensions on organizational outcomes (Figure 1).d Adds the direct paths from skill-enhancing HR practices to voluntary turnover and operational outcomes.e Model fit compared with the previous model.fDeletes the direct paths from skill-enhancing HR practices to voluntary turnover.g Adds the direct paths from motivation-enhancing HR practices to both voluntary turnover and operational outcomes.h Deletes the direct path from motivation-enhancing HR practices to operational outcomes.i Adds the direct path from opportunity-enhancing HR practices to voluntary turnover and operational outcomes.j Adds the direct path from HPWS to both voluntary turnover and operational outcomes.k Model fit compared with the theoretical model of the effects of HPWS on organizational outcomes (Figure 3).

*p .05**p .01



12/31

path from motivation-enhancing HR practices andoperational outcomes was not significant ( .02, n.s.), and we dropped it without impacting fit(model 5: 2[1] 0.76, n.s.). Finally, we added thedirect paths from opportunity-enhancing HR prac-tices to voluntary turnover and operational out-comes, and both paths were significant (model 6:

2

[2] 50.22, p .01). Therefore, we kept model6 as the final model for the mediation analyses.Figure 2 presents the standardized path estimates

for the final mediating model. Both human capitaland employee motivation were negatively relatedto voluntary turnover ( .20,p .01 for humancapital; .34,p .01 for employee motivation)and were positively related to operational out-comes ( .15,p .01 for human capital; .26,p .01 for employee motivation). In turn, volun-tary turnover was negatively related to financialoutcomes ( .08, p .01), whereas operationaloutcomes were positively associated with financialoutcomes ( .42, p .01). Sobel (1982) testsshowed that the indirect relationships between allthree HR dimensions and voluntary turnover, op-erational outcomes, and financial outcomes weresignificant (Zvaried from 8.05 to 13.89, allp-valueswere less than .01). In sum, these results suggest

that human capital, employee motivation, volun-tary turnover, and operational outcomes partiallymediated the relationships between skill-enhanc-ing and motivation-enhancing HR dimensions andfinancial outcomes and fully mediated the relation-ship between opportunity-enhancing HR practicesand financial outcomes. Hypotheses 5 through 7

were generally supported.We obtained the indirect effects and total effects

of the three HR dimensions on financial outcomesfrom the estimates in SEM. The total effects ofskill-enhancing, motivation-enhancing, and oppor-tunity-enhancing HR dimensions on financial out-comes were .13, 18, and .09 respectively (allps .01). The indirect effects mediated by human capi-tal, employee motivation, voluntary turnover, andoperational outcomes were .08, 05, and .09 for thethree HR dimensions respectively. We also calcu-lated the squared multiple correlations (i.e., R2s) for

structural equations predicting human capital (.22),employee motivation (.25), voluntary turnover(.18), operational outcomes (.22), and financial out-comes (.26). The results indicate that the finalmodel explained a moderate amount of variance inthese variables.

FIGURE 2Final Model of Effects of HR Dimensions on Organizational Outcomesa

HumanCapitalR2= .22

EmployeeMotivation

R2= .25

VoluntaryTurnoverR2= .18

OperationalOutcomesR2= .22

Skill-Enhancing

HR Practices

Motivation-Enhancing HR

Practices

Opportunity-

Enhancing HR

Practices

FinancialOutcomesR2= .26

.05**

.29**

.07**

.21**

.29**

.13**

.07**

.25**

.20**

.16**

.34**

.26**

.08**

.43**

.12**

.07**

.11**

.05**

.46**

.44**

.47**

a Standardized coefficients are presented; n 3,714.** p .01



13/31

In addition, we conducted a post hoc analysis toexamine whether the three-dimensional model(model 6) fit the data better than a unidimensionalmodel that treats the three HR dimensions as indi-cators of high-performance work systems (HPWS;Figure 3). As shown in Table 3, the partial mediat-ing model (model 7), in which HPWS has direct

impact on voluntary turnover, operational out-comes, and financial outcomes, fit the data well(2[14] 406.51, RMSEA .09, CFI .96, SRMR .03). Because the two models (6 and 7) were notnested, we relied on indexes other than chi-squarechange to compare them. In general, the three-di-mensional model (model 6: RMSEA .08, CFI .99, SRMR .02) fit better than unidimensionalmodel 7, but the differences in fit indexes were notgreat. Then we used an additional fit index, Akai-kes information criterion (AIC; Akaike, 1974),which is generally used in SEM to compare non-nested models estimated with the same data (Hen-son, Reise, & Kim, 2007; Kline, 2005). The value ofAIC itself does not indicate the quality of a model;only the AIC relative to that of another model ismeaningful. Lower values indicate a better fit, andso the model with the lowest AIC is the best fittingone. As shown in Table 3, the AIC for model 6(190.32) was lower than that for model 7 (450.51),which indicates the three-dimensional model fitthe data better than the unidimensional model.

DISCUSSION

Our aim in this meta-analytic review is to con-tribute to strategic HRM research by exploring the

mediating mechanisms through which HR prac-tices influence organizational outcomes. Drawingupon the ability-motivation-opportunity model ofHRM, the behavioral perspective of HRM, humancapital theory, and the resource-based view of thefirm, we proposed and found that the three dimen-sions of HR systems had differential relationships

with human capital and employee motivation,which were in turn related to voluntary turnoverand operational outcomes, and were further asso-ciated with financial outcomes. In addition, ourfindings demonstrated the direct relationships be-tween skill-enhancing HR practices and motiva-tion-enhancing HR practices and financial out-comes. Below we discuss the research and practicalimplications of our findings.

Research Implications

This research offers a number of important theo-retical contributions. First, we adopt multiple the-oretical perspectives on HRM to extend previousmediating models of HRMs influence on organiza-tional outcomes (e.g., Becker & Huselid, 1998; Del-ery & Shaw, 2001; Guest, 1997). Drawing upon thebehavioral perspective on HRM, human capital the-ory, and the resource-based view, the current studydemonstrates that HRM positively relates to finan-cial performance both by encouraging desired em-ployee behaviors and by building a valuable humancapital pool. It also suggests that future researchshould simultaneously address the mediating roles

of human capital and employee motivation so thatit can provide a clearer understanding of the link-

FIGURE 3Theoretical Model of Effects of HPWS on Organizational Outcomes

Skill-Enhancing

HR Practices

Motivation-

Enhancing HR

Practices

Opportunity-

Enhancing HR

Practices

Human

Capital

Employee

Motivation

Voluntary

Turnover

Operational

Outcomes

Financial

Outcomes

HighPerformance

Work Systems



14/31

age between HRM and operational and financialoutcomes.

Moreover, this study embraced the multidimen-sionality of performance as well as the potential for

different relationships with proximal and distaloutcomes. Researchers have recently called forstudies to simultaneously examine multiple out-come variables that have only been studied inde-pendently before (Lengnick-Hall et al., 2009). Withthe help of meta-analytic techniques, we tested acomprehensive mediating model and provided em-pirical support for the theoretical proposition thatHRM first relates to proximal outcomes, which fur-ther relate to distal outcomes (Becker & Huselid,1998; Delery & Shaw, 2001; Dyer & Reeves, 1995;Guest, 1997) and revealed that the relationships

between HRM and distal outcomes (e.g., opera-tional and financial outcomes) could be mediatedthrough multiple pathways (e.g., through humancapital and employee motivation). Moreover, as weexpected, there were direct relationships betweenskill-enhancing HR practices and motivation-en-hancing HR practices and financial outcomes thatcould not be explained by the mediating process.This is consistent with prior research suggestingthat HRM can improve organizational effectivesthrough alternative approaches such as affectinginternal interaction within organizations (Evans &

Davis, 2005; Gittell et al., 2010) and enhancing thesocial capital of organizations (Collins & Clark,2003). The findings of the current study and otherssuggest that it is meaningful for future research to

further explore other mediators of the relationshipbetween HRM and organizational outcomes.

One major contribution of this study to the stra-tegic HRM literature is that the results suggest dif-ferential effects of the three dimensions of HR sys-tems. This finding is important both in theory andin the methodology of measuring HR systems. The-oretically, this finding challenges previous re-search, in which the assumption has been that allHR practices in an HR system function in the samepattern. Our findings remind researchers that dif-ferent dimensions of HR systems may have unique

relationships with specific organizational out-comes. For example, skill-enhancing HR practiceswere more effective in enhancing human capital,whereas motivation-enhancing HR practices andopportunity-enhancing HR practices were morelikely to improve employee motivation. This resultis also consistent with recent research suggestingthe heterogeneous effects of the components of HRsystems on organizational outcomes (e.g., Batt &Colvin, 2011; Gardner et al., 2011; Gong et al., 2009;Liao et al., 2009; Shaw et al., 2009; Subramony,2009). HR practices are not only distinct, but also

FIGURE 4Effects of HPWS on Organizational Outcomesa

.67**

Skill-Enhancing

HR Practices

Motivation-

Enhancing HR

Practices

Opportunity-

Enhancing HR

Practices

HumanCapitalR2= .34

EmployeeMotivation

R2= .38

VoluntaryTurnoverR2= .15

OperationalOutcomes

R

2

= .22

FinancialOutcomesR2= .27

HighPerformance

Work Systems

.67**

.64**

.58**

.62**

.21**

.10**

.03

.28*

.16**

.05**

.37**

.08*

.34**a Standardized coefficients are presented; n 3,714.

* p .05** p .01



15/31

operate via different pathways. Therefore, we en-courage additional research to explore the influ-ence of these components of HR systems to advanceknowledge of the relationship between HRM andorganizational outcomes.

The findings of the differential relationships be-tween the dimensions of HR systems and organiza-

tional outcomes also offer methodological implica-tions for strategic HRM research. First, if all threedimensions of HR systems have unique effects onorganizational outcomes, failure to include any di-mension may compromise the overall impact of HRsystems on organizational outcomes or at least leadto inaccurate results. Moving forward, we encour-age researchers to include all three HR dimensionsin their measures of HR systems. Moreover, theresults show that the three HR dimensions havedifferential relationships with human capital andemployee motivation. Relatedly, the results indi-cate that the three-dimensional model fit the dataslightly better than the model combining the threeHR dimensions into a unidimensional HPWS ele-ment. Combined, these findings offer preliminaryevidence that the three HR dimensions are betterviewed as three distinct but related components ofHR systems rather than interchangeable indicatorsof HR systems. This suggestion is consistent withprevious research that argued that the measure ofHR systems should be formative rather than reflec-tive (e.g., Jiang, et al., 2012; Shaw et al., 2005,2009), and it encourages researchers to reconsiderwhether it is appropriate to utilize addition of HR

practices to represent HR systems. As an alternativeapproach, researchers might categorize HR prac-tices into the three HR dimensions and exploretheir main effects and interactions on organiza-tional outcomes (e.g., Gardner et al., 2011). In ad-dition, we encourage future research to comparethe use of multidimensional and unidimensionalmodels of HR systems and their effects on organi-zational outcomes. This stream of research can fur-ther verify the findings of this study and offer im-plications for the measurement of HR systems.

Practical Implications

Our study also offers implications for managerialpractices. First of all, our finding indicates that theinvestment in three HR dimensions was associatedwith the increase in financial outcomes. Specifi-cally, we found that given no change in other con-ditions, a one standard deviation increase in skill-enhancing, motivation-enhancing, or opportunity-enhancing HR practices was related to a .13, .18, or.09 standard deviation increase in financial out-comes. For example, Huselid (1995) examined the

relationship between motivation-enhancing HRpractices and financial performance and reported amean and standard deviation of 0.46 and 1.64 forTobinsQ. If we apply our finding to this study, onestandard deviation increase in motivation-enhanc-ing HR practices is associated with 64 percent im-provement in Tobins Q. This result suggests that

organizations can obtain substantial financial ben-efits from investing in the three HR dimensionsconsidered here.

In addition, the results of this study shed light onthe ways through which managers can increase thebenefits of investing in HRM. The results indicatethat to retain talented employees and realize oper-ational and financial objectives, organizations needto use HR practices to enhance both employeeskills and motivation at work. More specifically,we suggest that organizations focus more on prac-tices, such as recruitment, selection, and training

when enhancing employee skills. In contrast, whenorganizations aim to improve employee motiva-tion, they should consider how to appraise employ-ees performance, how to compensate for theirwork, how to make jobs meaningful and interest-ing, and how to involve employees in work teamsand decision making. With these suggestions, how-ever, we do not deny the potential effects of recruit-ment, selection, and training in enhancing em-ployee motivation or the positive impact ofperformance appraisal, compensation, job design,or employee involvement in developing employ-

ees human capital. Instead, we encourage organi-zations to maximize the return on their investmentin HRM by using appropriate HR practices. Forexample, in order to improve employee motivation,it may be wise to check whether performance ap-praisal and compensation systems appropriatelyreflect employees contribution at work rather thantraining employees how to complete their work.

Our study also indicates that organizations in-vestment in HRM leads to financial outcomesthrough a mediating process. Any other factors thatcan impact the intermediate variables may affectthe effects of HRM on the distal financial outcomes.This reminds managers of attending to whethertheir HR practices improve employee skills andmotivation effectively and whether other manage-rial initiatives can boost or undermine the effects ofHR practices. For example, researchers have re-ported that leadership and organizational culturehave an important impact on employee motivation(Hartnell, Ou, & Kinicki, 2011; Ilies, Nahrgang, &Morgeson, 2007). Therefore, managers may con-sider how these factors can complement the effectsof HR practices in enhancing employee motivation.



16/31

Limitations and Future Research

Several limitations should be noted in the cur-rent study. First, some studies included in thismeta-analysis used informant-reported measures toevaluate HR practices and organizational outcomesfrom the same source. This may lead to commonmethod bias, which might inflate the correlationsbetween HR practices and organizational out-comes. Relatedly, most of the studies included inthe analysis had cross-sectional designs, whichmay limit conclusions regarding the direction ofthe mediating mechanism. The results from thecurrent investigation should be interpreted withthese limitations in mind. We encourage more lon-gitudinal studies that collect information on HRpractices and organizational outcomes from differ-ent sources. Future meta-analysis can explore if alongitudinal research design may influence the es-timates of effect sizes and the mediating mecha-nisms examined in this study.

Second, potential moderators may exist in therelationships among HR dimensions and organiza-tional outcome categories. For example, recentmeta-analytic reviews have reported that industrytype (manufacturing industry vs. service industry)and country moderated the relationship betweenHPWS and organizational outcomes (Combs et al.,2006; Rabl, Jayasinghe, Gerhart, & Kuehlmann,2011; Subramony, 2009). Researchers have alsosuggested that HR practices applied to a specificgroup of employees, or used for employees in gen-eral, may influence their effects on organizationaloutcomes (Gerhart, Wright, McMahan, & Snell,2000). However, owing to the relatively few studiesin the subgroups divided by the potential modera-tors, we were not able to test the mediating modelseparately in each subgroup. Future research canexamine this mediating model by using samplesfrom different industries, different countries, anddifferent job groups.

A third limitation of this study is that we wereunable to explore synergy among the three HR di-mensions by examining their interactions, eventhough the synergies within HR systems have beensuggested in the literature (e.g., Delery, 1998; Ger-hart, 2007; Jiang et al., 2012). Operationally, thiswas impossible owing to how existing studies weremeasured. However, moving forward, if a goodamount of research includes all three HR dimen-sions while reporting the correlations of HR dimen-sions and organizational outcomes with interactionterms comprised of the three HR dimensions, fu-ture meta-analytic review will be able to exam-ine this.

Fourth, in the current study we examined volun-tary turnover as an intermediate outcome mediat-ing the relationships between the three HR dimen-sions as well as employee human capital andmotivation and financial outcomes. However, agrowing literature indicates that voluntary turnovermay moderate the relationship between HRM and

financial outcomes (e.g., Guthrie, 2001; Haus-knecht & Trevor, 2011; Shaw, 2011; Shaw et al.,2005). However, we were not able to test the inter-actions between the three HR dimensions and vol-untary turnover because very few studies reportedthe correlations between the interaction terms andthe variables examined. We encourage scholars toexplore this issue in future research. In addition,recent turnover research suggests that involuntaryturnover or dismissal is also influenced by HRpractices and negatively related to operational andfinancial outcomes (Batt & Colvin, 2011; Haus-knecht & Trevor, 2011). It is worth considering theroles of both types of turnover in the mediatingprocess rather than just focusing on voluntaryturnover.

Fifth, like other meta-analyses testing mediatingprocess (e.g., Chang, Rosen, & Levy, 2009; Colquitt,Scott, & LePine, 2007; Robbins, Oh, Le, & Button,2009), the current meta-analysis did not includecontrol variables in the regression models (e.g., in-dustry, size, unionization, strategy) because manystudies did not provide correlations with thesevariables.

Finally, our study only focused on the relation-

ships between HRM and organizational outcomesat the organizational level, even though there is agrowing research focus on cross-level influences oforganization-level HRM on individual-level out-comes (e.g., Liao et al., 2009; Snape & Redman,2010; Takeuchi et al., 2009) and on the influence ofemployee-perceived HR systems on individual out-comes (e.g., Butts, Vandenberg, DeJoy, Schaffer, &Wilson, 2009; Kehoe & Wright, in press). We en-courage more empirical studies on the effects oforganization-level HR systems and employee-per-ceived HR systems on individual outcomes. Over

time, there may be enough studies for a futuremeta-analysis summarizing these effects on indi-vidual outcomes.

Conclusions

This meta-analysis examined and extended thetheoretical model linking human resource manage-ment with organizational outcomes (e.g., Becker &Huselid, 1998; Delery & Shaw, 2001; Guest, 1997).We found that three dimensions of HR systems (i.e.,skill-enhancing, motivation-enhancing, and oppor-



17/31

tunity-enhancing HR practices) were positively re-lated to human capital and employee motivation indifferent patterns in such a way that, comparedwith the other two HR dimensions, skill-enhancingHR practices were more positively related to hu-man capital and less positively related to employeemotivation. In addition, human capital and em-

ployee motivation mediated the relationships be-tween three HR dimensions and voluntary turnoverand operational outcomes, which in turn related tofinancial outcomes. We also found direct relation-ships between the three dimensions of HR systemsand voluntary turnover, operational outcomes, andfinancial outcomes and thus encourage future re-search exploration of additional mediators in therelationships between HRM and organizationaloutcomes.

REFERENCESa

Agarwala, T. 2003. Innovative human resource practicesand organizational commitment: An empirical in-vestigation. International Journal of Human Re-source Management,14: 175197.

Aguinis, H., Pierce, C. A., Bosco, F. A., Dalton, D. R., &Dalton, C. M. 2011. Debunking myths and urbanlegends about meta-analysis. Organizational Re-search Methods,14: 306331.

*Ahmad, S., & Schroeder, R. G. 2003. The impact ofhuman resource management practices on opera-tional performance: Recognizing country and indus-try differences. Journal of Operations Manage-

ment,21: 1943.Akaike, H. 1974. A new look at the statistical model

identification. IEEE Transactions on AutomaticControl,19: 716723.

*Akhtar, S., Ding, D. Z., & Ge, G. L. 2008. Strategic HRMpractices and their impact on company performancein Chinese enterprises. Human Resource Manage-

ment,47: 1532.

Appelbaum, E., Bailey, T., Berg, P., & Kalleberg, A. L.2000.Manufacturing advantage: Why high-perfor-

mance work systems pay off.Ithaca, NY: CornellUniversity Press.

*Appleyard, M. M., & Brown, C. 2001. Employment prac-tices and semiconductor manufacturing perfor-mance.Industrial Relations,40: 436471.

*Armstrong, C., Flood, P. C., Guthrie, J. P., Liu, W., Mac-Curtain, S., & Mkamwa, T. 2010. Beyond high perfor-mance work systems: The impact of including diver-sity and equality management on firm performance.

Human Resource Management,49: 977998.

*Arthur, J. B. 1994. Effects of human resource systems onmanufacturing performance and turnover.Academyof Management Journal,37: 670687.

*Audea, T., Teo, S. T. T., & Crawford, J. 2005. HRMprofessionals and their perceptions of HRM and firmperformance in the Philippines.International Jour-

nal of Human Resource Management,16: 532552.

Bae, J., & Lawler, J. J. 2000. Organizational and HRMstrategies in Korea: Impact on firm performance inan emerging economy. Academy of Management

Journal,43: 502517.

Bailey, T. 1993. Discretionary effort and the organiza-tion of work: Employee participation and work

reform since Hawthorne.Paper prepared at Colum-bia University for the Sloan Foundation.

Bailey, T., Berg, P., & Sandy, C. 2001. The effect of highperformance work practices on employee earnings inthe steel, apparel, and medical electronics and im-aging industries. Industrial & Labor Relations Re-

view,54: 525543.*Barksdale, W. K. 1994. Human resource management,

organizational climate, work attitudes and organ-izational performance.Unpublished doctoral dis-sertation, Georgia State University.

Barling, J., Kelloway, E. K., & Iverson, R. D. 2003. High-quality work, job satisfaction, and occupational in-juries.Journal of Applied Psychology,88: 276283.

Barney, J. 1991. Firm resources and sustained competi-tive advantage. Journal of Management, 17: 99120.

*Bartram, T., Stanton, P., Leggat, S., Casimir, G., & Fraser,B. 2007. Lost in translation: Exploring the link be-tween HRM and performance in healthcare. Human

Resource Management Journal,17: 2141.

*Batt, R. 2002. Managing customer services: Human re-source practices, quit rates, and sales growth.Acad-emy of Management Journal,45: 587597.

*Batt, R., & Colvin, A. J. S. 2011. An employment systemsapproach to turnover: HR practices, quits, dismiss-als, and performance. Academy of Management

Journal,54: 695717.

*Batt, R., Colvin, A. J. S., & Keefe, J. 2002. Employee

voice, human resource practices, and quit rates: Ev-idence from the telecommunications industry. In-dustrial & Labor Relations Review,55: 573594.

Becker, B. E., & Gerhart, B. 1996. The impact of humanresource management on organization performance:Progress and prospects. Academy of Management

Journal,39: 779801.

Becker, B. E., & Huselid, M. A. 1998. High performancework systems and firm performance: A synthesis ofresearch and managerial implications. In G. R. Ferris(Ed.),Research in personnel and human resources

management:53101. Greenwich, CT: JAI Press.

a Articles used in the meta-analysis are marked withan asterisk.



18/31

Becker, G. S. 1964. Human capital: A theoretical andempirical analysis.Chicago: University of ChicagoPress.

*Beltrn-Martn, I., Roca-Puig, V., Escrig-Tena, A., &Bou-Llusar, J. 2008. Human resource flexibility as amediating variable between high performance worksystems and performance.Journal of Management,

34: 10091044.Blau, P. 1964. Exchange and power in social life. New

York: Wiley.

Boxall, P., & Purcell, J. 2008. Strategy and human re-source management. Basingstoke, U.K.: PalgraveMacmillan.

*Brown, M. P., Sturman, M. C., & Simmering, M. J. 2003.Compensation policy and organizational perfor-mance: The efficiency, operational, and financial im-plications of pay levels and pay structure. Academyof Management Journal,46: 752762.

Butts, M. M., Vandenberg, R. J., DeJoy, D. M., Schaffer,B. S., & Wilson, M. G. 2009. Individual reactions tohigh involvement work processes: Investigating therole of empowerment and perceived organizationalsupport. Journal of Occupational Health Psychol-ogy,14: 122136.

*Cabello-Medina, C., Lopez-Cabrales, A., & Valle-Ca-brera, R. 2011. Leveraging the innovative perfor-mance of human capital through HRM and socialcapital in Spanish firms. International Journal of

Human Resource Management,22: 807828.

Campbell, J. P., McCloy, R. A., Oppler, S. H., & Sager,

C. E. 1993. A theory of performance. In N. Schmitt &W. C. Borman (Eds.), Personnel selection in organ-

izations:3570. San Francisco: Jossey-Bass.

Capon, N., Farley, J. U., & Hoenig, S. 1990. Determinantsof financial performance: A meta-analysis.Manage-

ment Science,36: 11431159.

Cappelli, P., & Neumark, D. 2001. Do high-performancework practices improve establishment-level out-comes?Industrial and Labor Relations Review,54:737775.

*Chan, L. L. M., Shaffer, M. A., & Snape, E. 2004. Insearch of sustained competitive advantage: The im-

pact of organizational culture, competitive strategyand human resource management practices on firmperformance. International Journal of Human Re-source Management,15: 1735.

*Chandler, G. N., & McEvoy, G. M. 2000. Human re-source management, TQM, and firm performance insmall and medium-size enterprises.Entrepreneur-ship: Theory and Practice,25: 4357.

Chang, C., Rosen, C. C., & Levy, P. E. 2009. The relation-ship between perceptions of organizational politicsand employee attitudes, strain, and behavior: A

meta-analytic examination. Academy of Manage-ment Journal,52: 779801.

*Chen, C., & Huang, J. 2009. Strategic human resourcepractices and innovation performanceThe mediat-ing role of knowledge management capacity.Journalof Business Research,62: 104114.

Cheung, M. W. L., & Chan, W. 2009. A two-stage ap-

proach to synthesizing covariance matrices in meta-analytic structural equation modeling. Structural

Equation Modeling,16: 2853.

Cheung, M. W. L., & Chan, W. 2005. Meta-analytic struc-tural equation modeling: A two-stage approach.Psy-chological Methods,10: 4064.

*Chuang, C., & Liao, H. 2010. Strategic human resourcemanagement in service context: Taking care of busi-ness by taking care of employees and customers.

Personal Psychology,63: 153196.

Clogg, C. C., Petkova, E., & Haritou, A. 1995. Statisticalmethods for comparing regression coefficients be-tween models.American Journal of Sociology,100:12611293.

Coff, R. W. 1997. Human assets and management dilem-mas: Coping with hazards on the road to resource-based theory. Academy of Management Review,22: 374402.

Coff, R. W. 2002. Human capital, shared expertise, andthe likelihood of impasse in corporate acquisitions.

Journal of Management,28: 107128.

Collins, C. J., & Clark, K. D. 2003. Strategic human re-source practices, top management team social net-works, and firm performance: The role of humanresource practices in creating organizational compet-itive advantage.Academy of Management Journal,46: 740751.

*Collins, C. J., & Smith, K. G. 2006. Knowledge exchangeand combination: The role of human resource prac-tices in the performance of high-technology firms.

Academy of Management Journal,49: 544560.

*Collins, C. J., Smith, K. G., & Stevens, C. K. 2001. Hu-man resource practices, knowledge-creation capa-bility and performance in high technology firms.CAHRS working paper, Center for Advanced HumanResource Studies, Cornell University.

Colquitt, J. A., Scott, B. A., & LePine, J. A. 2007. Trust,trustworthiness, and trust propensity: A meta-ana-lytic test of their unique relationships with risk tak-ing and job performance. Journal of Applied Psy-chology,92: 909927.

*Colvin, A. J. S., Batt, R., & Keefe, J.2005.The impact ofemployee voice and compliance mechanisms onabsenteeism, discipline, and turnover. CAHRSworking paper, no. 05-13, Center for Advanced Hu-man Resource Studies, Cornell University.

Combs, J., Liu, Y., Hall, A., & Ketchen, D. 2006. How



19/31

much do high-performance work practices matter? Ameta-analysis of their effects on organizational per-formance.Personnel Psychology,59: 501528.

Crook, T. R., Todd, S. Y., Combs, J. G., Woehr, D. J., &Ketchen, D. J. 2011. Does human capital matter? Ameta-analysis of the relationship between humancapital and firm performance. Journal of Applied

Psychology,96: 443456.Curtis, B., Hefley, W. E., & Miller, S. 1995.People capa-

bility maturity model. Pittsburgh: Pittsburgh Soft-ware Engineering Institute, Carnegie Mellon Univer-sity.

Dalton, D. R., Daily, C. M., Certo, S. T., & Roengpitya, R.2003. Meta-analyses of financial performance andequity: Fusion or confusion? Academy of Manage-

ment Journal,46: 1326.

Dalton, D. R., Daily, C. M., Ellstrand, A. E., & Johnson,J. L. 1998. Meta-analytic reviews of board composi-tion, leadership structure, and financial perfor-

mance. Strategic Management Journal, 19: 269290.

Dalton, D. R., Daily, C. M., Johnson, J. L., & Ellstrand,A. E. 1999. Number of directors and financial per-formance: A meta-analysis. Academy of Manage-

ment Journal,42: 674686.

Dalton, D. R., & Todor, W. D. 1979. Turnover turned over:An expanded and positive perspective.Academy of

Management Review,4: 225236.

*Datta, D. K., Guthrie, J. P., & Wright, P. M. 2005. Humanresource management and labor productivity: Doesindustry matter?Academy of Management Journal,

48: 135145.

*De Winne, S., & Sels, L. 2010. Interrelationships be-tween human capital, HRM and innovation in Bel-gian start-ups aiming at an innovation strategy. In-ternational Journal of Human Resource

Management,21: 18631883.

Deci, E. L., Connell, J. P., & Ryan, R. M. 1989. Self-determination in a work organization. Journal of

Applied Psychology,74: 580590.

*Delaney, J. T., & Huselid, M. A. 1996. The impact ofhuman resource management practices on percep-tions of organizational performance. Academy of

Management Journal,39: 949969.

Delery, J. E. 1998. Issues of fit in strategic human re-source management: Implications for research. Hu-

man Resource Management Review, 8: 289309.

*Delery, J. E., & Doty, D. H. 1996. Modes of theorizing instrategic human resource management: Tests of uni-versalistic, contingency, and configurational perfor-mance predictions.Academy of Management Jour-

nal,39: 802835.

*Delery, J. E., Gupta, N., Shaw, J. D., Douglas Jenkins,J. G., & Ganster, M. L. 2000. Unionization, compen-

sation, and voice effects on quits and retention. In-dustrial Relations,39: 625645.

Delery, J. E., & Shaw, J. D. 2001. The strategic manage-ment of people in work organizations: Review, syn-thesis, and extension. In G. R. Ferris (Ed.), Research

in personnel and human resource management,vol. 20: 167197. Stamford, CT: JAI.

*Den Hartog, D. N., & Verburg, R. M. 2004. High perfor-mance work systems, organizational culture andfirm effectiveness. Human Resource Management

Journal,14: 5578.

Dess, G. G., & Shaw, J. D. 2001. Voluntary turnover, socialcapital, and organizational performance. Academyof Management Review,26: 446456.

Dyer, L., & Reeves, T. 1995. Human resource strategiesand firm performance: What do we know and wheredo we need to go? International Journal of Human

Resource Management,6: 656 670.

Erez, A., Bloom, M. C., & Wells, M. T. 1996. Using ran-dom rather than fixed effects models in meta-analy-sis: Implications for situational specificity and valid-ity generalization. Personnel Psychology,49: 275306.

*Ericksen, J. 2006.High-performance work systems, dy-namic workforce alignment, and firm perfor-mance.Unpublished doctoral dissertation, CornellUniversity.

Evans, W. R., & Davis, W. D. 2005. High-performancework systems and organizational performance: Themediating role of internal social structure.Journal of

Management,31: 758775.

*Faems, D., Sels, L., De Winne, S., & Maes, J. 2005. Theeffect of individual HR domains on financial perfor-mance: Evidence from Belgian small businesses. In-ternational Journal of Human Resource Manage-

ment,16: 676700.

*Fey, C. F., & Bjrkman, I. 2001. The effect of humanresource management practices on MNC subsidiaryperformance in Russia. Journal of International

Business Studies,32: 5975.

*Fey, C. F., Bjrkman, I., & Pavlovskaya, A. 2000. Theeffect of human resource management practices onfirm performance in Russia. International Journal

of Human Resource Management,11: 118.

*Gardner, T. M., Wright, P. M., & Moynihan, L. M. 2011.The impact of motivation, empowerment, and skill-enhancing practices on aggregate voluntary turn-over: The mediating effect of collective affectivecommitment.Personal Psychology,64: 315350.

*Gelade, G. A., & Ivery, M. 2003. The impact of humanresource management and work climate on organi-zational performance. Personnel Psychology, 56:383404.

Gerhart, B. 2007. Horizontal and vertical fit in human



20/31

resource systems. In C. Ostroff & T. A. Judge (Eds.),Perspectives on organizational fit:317348. NewYork: Psychology Press.

*Gerhart, B., & Milkovich, G. T. 1990. Organizationaldifferences in managerial compensation and finan-cial performance. Academy of Management Jour-

nal,33: 663691.

Gerhart, B., Wright, P., McMahan, G., & Snell, S. 2000.Research on human resource decisions and firm per-formance: How much error is there, and how does itinfluence effect size estimates? Personnel Psychol-ogy,53: 803

Documents

Meta Analytic Analysis of SHRM