The Leadership Quarterly 20 (2009) 371–382

Contents lists available at ScienceDirect

The Leadership Quarterly

j ourna l homepage: www.e lsev ie r.com/ locate / l eaqua

Leader–member exchange and empowerment: Direct and interactive effectson job satisfaction, turnover intentions, and performance

Kenneth J. Harris a,⁎, Anthony R. Wheeler b,1, K. Michele Kacmar c,2

a Indiana University Southeast, School of Business, 4201 Grant Line Road, New Albany, IN 47150, United Statesb Schmidt Labor Research Center, College of Business Administration, University of Rhode Island, 36 Upper College Road, Kingston, RI 02881, United Statesc Durr-Fillauer Chair of Business Ethics, Department of Management and Marketing, Culverhouse College of Commerce and Business Administration,143 Alston Hall, Box 870225, The University of Alabama, Tuscaloosa, Alabama 35487-0225, United States

a r t i c l e i n f o

⁎ Corresponding author. Tel.: +1 812 941 2501.E-mail addresses: harriskj@ius.edu (K.J. Harris), ar

1 Tel.: +1 401 874 9491.2 Tel.: +1 205 348 8931.

1048-9843/$ – see front matter © 2009 Elsevier Inc.doi:10.1016/j.leaqua.2009.03.006

a b s t r a c t

Keywords:

This study examined the moderating impact of empowerment on the relationships betweenleader–member exchange (LMX) quality and the self-rated outcomes of job satisfaction andturnover intentions, as well as the supervisor-rated outcomes of job performance andorganizational citizenship behaviors. Two samples, with 244 and 158 employees respectively,were used to test our hypotheses. Our results provided evidence that in general, empowermentmoderates the relationships between LMX and job outcomes. These findings are important asprevious research has only tested these variables as independent predictors, but our resultssuggest the relationships these constructs havewith important consequences are dependent onboth variables. Practical implications and directions for future research are offered.

© 2009 Elsevier Inc. All rights reserved.

LMXEmpowermentJob satisfactionTurnover intentionsJob performance

1. The interactive effects of leader–member exchange and empowerment on job satisfaction, turnover intentions, andperformance

The quality of relationships between supervisors and subordinates, often studied via leader–member exchange (LMX) theory,has been the focus of considerable research attention (e.g., Graen, 2004; Graen & Uhl-Bien, 1995). The results of this research haveshown LMX to be positively related to desired outcomes including increased job performance (Gerstner & Day, 1997; Schriesheim,Castro, & Cogliser, 1999), contextual performance (e.g. Ilies, Nahrgang, & Morgeson, 2007), motivation (Tierney, Farmer, & Graen,1999), job satisfaction (Gerstner & Day, 1997; Epitropaki & Martin, 2005), and organizational commitment (Martin, Thomas,Charles, Epitropaki, & McNamara, 2005; Schriesheim et al., 1999). The preponderance of empirical evidence surrounding thequality of the supervisor–subordinate relationship has led researchers to conclude that this relationship is one of the mostimportant an employee has (Liden, Sparrowe, & Wayne, 1997), and potentially one of the most important predictors of workplaceoutcomes (Manzoni & Barsoux, 2002).

Similar to LMX, empowerment programs at work have also received considerable attention from academics and practitionersalike. Psychological empowerment (Spreitzer, 1995), describes the increase in task motivation through the influence of fourcognitions: meaning, competence, impact, and self-determination. These cognitions reflect employees' orientation toward theirjobs and have been associated with positive results (e.g., Spreitzer, 1995; Spreitzer, Kizilos, & Nason, 1997).

Thus, empirical evidence has found both LMX and empowerment to be positively related to important organizationalbehaviors. Further, although early behavioral studies of leadership found that effective leaders empowered employees tomake job-

wheeler@mail.uri.edu (A.R. Wheeler), mkacmar@ua.edu (K.M. Kacmar).

All rights reserved.

372 K.J. Harris et al. / The Leadership Quarterly 20 (2009) 371–382

related decisions (Likert, 1961), empowerment and LMX are fundamentally different constructs. LMX describes the quality of thesupervisor–subordinate relationship, with qualities varying from low to high (e.g., Liden, Erdogan, Wayne, & Sparrowe, 2006),whereas empowerment relates to employee motivation towards work. As it would seem that studies including LMX andempowerment would be common, surprisingly, few research efforts have examined these constructs in the same study (e.g., Liden,Wayne, & Sparrowe, 2000). In one of the few studies that examined these constructs, Liden et al. (2000) found that although bothconstructs independently predicted employee behavioral consequences, researchwas needed that examined the interactive effectsof these constructs. Additionally, previous researchers have called for studies investigating moderators of the LMX–outcomerelationships (Schriesheim et al., 1999). It remains unclear in the literature if LMX and empowerment act independently, jointly, orcompetitively in their relation to critical employee behaviors.

As such, this study sets out to examine the moderating impact of empowerment on the relationships between LMX and theoutcomes of job satisfaction, turnover intentions, job performance, and organizational citizenship behaviors (OCBs). We do so byfirst theoretically developing our interaction hypotheses. Second, we use two samples to empirically test our hypotheses, and inthe process provide evidence related to the generalizability and/or boundary conditions of our results.

2. Leader–member exchange

The quality of relationships between supervisors and subordinates is often studied via LMX theory. With its roots in role theory(e.g., Kahn, Wolfe, Quinn, Snoek, & Rosenthal, 1964) and social exchange theory (Blau, 1964; Cropanzano & Mitchell, 2005;Homans, 1958), the LMX model suggests that supervisors form differential relationships with their subordinates. Theserelationships range in quality from high to low. Subordinates in high quality exchanges have relationships with their supervisorscharacterized by emotional support and trust (Dienesch & Liden, 1986). The benefits of high quality LMX relationships arenumerous, including preferential treatment, increased job-related communication, differential allocation of formal and informalrewards, ample access to supervisors, and increased performance-related feedback (e.g., Dienesch & Liden, 1986; Elicker, Levy, &Hall, 2006; Graen & Uhl-Bien, 1995). Conversely, subordinates in low quality LMX relationships often experience the exactopposite; supervisors provide limited emotional support and trust and the subordinates receive few, if any, benefits outside theemployment contract (Dienesch & Liden, 1986; Gerstner & Day, 1997).

These advantages for high quality LMX subordinates, which can be explained by social exchange theory, are likely to be relatedto positive outcomes. More explicitly, high quality relationships are associated with subordinates receiving increased access,communication, and rewards (Dienesch & Liden,1986), all of which are related to elevated job satisfaction and performance (in theforms of task and contextual performance), and decreased turnover intentions (Gerstner & Day, 1997). However, we believe thatthe positive impact of LMX on job outcomes is most salient depending upon employee feelings of empowerment. In particular, wesuggest that thosewho are less empoweredwill benefit more from high quality LMX relationships. Thus, in the next sectionwewillexplain the empowerment construct and how empowerment moderates LMX–outcome relationships.

3. Empowerment

As previously mentioned, empowerment in the workplace has been a popular idea for managers to implement and academicsto study. At its core, psychological empowerment relates to cognitions about one's work (Spreitzer, 1995). One of the theoreticalfoundations for predictions related to empowerment can be found in the job characteristics theory (JCT) (Hackman & Oldham,1976). JCT suggests that the specific characteristics of an employee's job, skill variety, task identity, task significance, autonomy, andjob feedback form an individual's motivating potential score, which predicts three critical psychological states, experiencedmeaningfulness of work, experienced responsibility for outcomes, and knowledge of results (e.g., Hackman & Oldham,1976). Thus,the characteristics of the job act as a motivator to perform, and employees are empowered and feel satisfied with their jobs whenthey perform (Champoux, 1991). From the JCT perspective, psychological empowerment links the critical psychological states. Ingeneral, when employees perceive high levels of empowerment, they aremotivated towards their jobs and are likely to experiencepositive accompanying consequences (e.g., Spreitzer et al., 1997). We suggest that the feelings associated with empowerment willnot only be directly associated with job outcomes, as has been shown in previous research, but also moderates relationshipsbetween LMX and outcomes. It is this moderating effect that we are interested in and will discuss in the following section.

3.1. The moderating effect of empowerment on LMX–outcome relationships

The extant literature has established that LMX quality (Gerstner & Day, 1997) and empowerment (e.g., Seibert, Silver, &Randolph, 2004) are both positively associated with desired outcomes. However, little is known about how these variablesinteract. In particular, does the absence or presence of empowerment on the job make the relationships between LMX and joboutcomes stronger or weaker? We suggest that the established relationships between LMX and the consequences of jobsatisfaction, turnover intentions, job performance, and OCBs are strongest when empowerment is lowest, as employees are notmotivated by the jobs. Thus, our study may help to specify the situations when LMX has either a greater or lesser impact on joboutcomes. This next step in furthering the LMX and empowerment research stream is important as previous researchers havecalled for examinations of the interactions of these variables and how they affect organizational consequences. Liden et al. (2000)examined these variables in the same study, although not their interactive effects, and called for future studies on “the integrationof empowerment with interpersonal relationships” (p. 414) in understanding individual consequences.

373K.J. Harris et al. / The Leadership Quarterly 20 (2009) 371–382

Based on social exchange and job characteristics theories, we expect to find the most positive outcomes when both LMX andempowerment are high, and the most negative outcomes when both were low. However, we expect LMX relationship quality willbe differentially important depending on how empowered an individual feels. The theories underlying LMX as well as JCT, thetheoretical foundation of empowerment, suggest that LMX relationship quality should be most important for those employeeswho perceive minimal empowerment from their jobs. The construct of empowerment explains how and why an employee's jobcan act as a motivator to energize, direct, and sustain the employee behaviors that ultimately are associated with both task andcontextual job performance and employee turnover intentions (e.g., Spreitzer, 1995). If an employee is high in empowerment andthus motivated by the job itself, the relationship with a supervisor is of less importance as the job provides the motivation andwhich is associated with more positive outcomes.

On the other hand, when empowerment is low, employees will have to look to other aspects of their work lives to find themotivation they do not receive from empowerment. The LMX relationship, which has been argued to be one of, if not the mostimportant relationship an employee has (e.g., Manzoni & Barsoux, 2002), becomes a logical place to look for benefits on the job.Social exchange theory leads us to believe that when the empowerment provided by the organization is lacking, social exchangesmay be of heightened importance. In our study, when an employee is not empowered and thus lacks the motivation related to his/her job, leaders are able to offer benefits that can keep employees motivated in a different way (Liden et al., 1997). Thus, socialexchange theory would lead us to believe that when empowerment was lacking, high quality exchanges would be especiallyimportant and be associated with employees still performing well, being satisfied, and wanting to stay with the organization.

Research on LMX theory presents clear empirical support for this contention. Tierney et al. (1999) found that LMX relationshipquality acted as a motivating factor for employees in the job context, which helps explain Gerstner & Day's (1997) meta-analyticfindings that LMX relationship quality positively relates to employee task and contextual performance, job satisfaction, andemployee turnover intentions. Tying empowerment and LMX relationship quality together, social exchange theory suggests thatthe exchange of resources between subordinates and supervisors will predict both positive and negative outcomes (Cropanzano &Mitchell, 2005).

Thus, it has been established that LMX is positively related to outcomes (e.g., Gerstner & Day, 1997), but this relationship is ofless importance when an employee experiences high empowerment. The empowerment and related motivation make the LMXrelationship less essential. As a result, we expect the positive associations between LMX and desired job outcomes to be less strongwhen empowerment is high. On the other hand, when empowerment is low and— thus notmotivating employees, the exchange ofresources found in the LMX relationship will become more important and act as the primary motivator (demotivator). That is, anemployee experiencing low levels of empowerment will lack the motivation to perform both task and beyond task duties, willbecome dissatisfied with the job, and will likely develop greater turnover intentions unless the LMX relationship quality with thesupervisor can offset the low level of empowerment. Further, when empowerment is lacking, the exchangewith the supervisor canfill this void and becomes all the more integral (Cropanzano & Mitchell, 2005). Thus, when empowerment is low, the associationsbetween LMX and job outcomes becomes stronger as the social exchanges with a direct leader will either offset the lack ofmotivation (high quality LMX), or in low quality exchanges, in effect exacerbate the low motivation from low empowerment. As aresult, we expect LMX to be positively related to desired job outcomes, and these relationships to be strongest whenempowerment is low. Based on these arguments, we offer the following hypotheses:

Hypothesis 1. Empowermentmoderates the positive relationship between LMX quality and job satisfaction, such that the positiverelationship is strongest when empowerment is lowest.

Hypothesis 2. Empowerment moderates the negative relationship between LMX quality and turnover intentions, such that thenegative relationship is strongest when empowerment is lowest.

Hypothesis 3. Empowerment moderates the positive relationship between LMX quality and job performance, such that thepositive relationship is strongest when empowerment is lowest.

Hypothesis 4. Empowerment moderates the positive relationship between LMX quality and organizational citizenship behaviors,such that the positive relationship is strongest when empowerment is lowest.

4. Method

In this study we examined two separate samples. The reason for this research design was we wanted to examine themoderating impact of empowerment on the relationships between LMX and an attitude (job satisfaction), a behavioral intention(turnover intentions), and behaviors (job performance andOCBs). Unfortunately, wewere unable to examine all of these outcomesin a single data collection effort, thus we examined data from two samples to test the study's hypotheses.

4.1. Sample 1: Participants and procedure

Surveys were mailed to 2000 alumni of a private Midwestern university. The alumni were chosen randomly from theuniversity's list of graduates over the past forty years. The survey's introductory page informed the potential subjects that theirresponses would be anonymous, that the purpose of the survey was to gather information on workplace relationships andattitudes, and that they could return their completed surveys in a prepaid envelope. Surveys were returned by 333 respondents

374 K.J. Harris et al. / The Leadership Quarterly 20 (2009) 371–382

(16.7% response rate). After receiving the surveys, we contacted the alumni office to seek information relating to any potentialsystematic demographical differences between respondents and non-respondents. We were informed that the demographiccharacteristics of non-respondents were similar in age, years since graduation, gender, types of jobs, and other characteristics toour respondents.

Of the surveys that were returned, 89 were not usable for reasons such as employees indicating that they worked in very smallbusinesses (often 5 or less people), they had beenwith their employers for a short period of time, theywere currently unemployed,or because they provided incomplete data. After eliminating these unusable surveys, the resulting usable sample size was 244respondents (12.2% response rate). Although somewhat low, this final usable response rate is consistent with previous studiesutilizing alumni survey research sampling (e.g., Seibert, Crant, & Kraimer, 1999). Further, our returned completed surveys wouldlikely have been higher, but wewere forced to rely onmailing lists from the university's alumni office and a considerable number ofsurveys were returned due to incorrect mailing addresses. Additionally, we were not able to target only employed alumni eventhough our study was only applicable to these individuals. Thus, our response rate does not take into account those individualswho were either involuntarily or voluntarily unemployed (Eby, Butts, & Lockwood, 2003).

This sample was composed of 49% women and 51% men. The average age of our respondents was 46.6 (SD=10.71) years andthe average organizational tenure was 12.82 (SD=9.97) years. Each of our respondents had received an undergraduate degree,with a large number having pursued graduate studies (i.e., law school, medical school, master's program, PhD). The occupationalprofile of our respondents varied considerably, with employees having jobs including teachers, accountants, bankers, engineers,professors, secretarial and administrative positions, front-line managers, middle managers, and CEOs.

4.2. Sample 2: Participants and procedure

Sample 2 was comprised of 158 (73% response rate) full-time employees at a state agency. These individuals worked on anumber of different aspects of environmental health, including radiation controls, chemical leakages, and water cleanliness.Employees at the state agency primarily completed their work individually, but were required to interact with their supervisors toeffectively complete all aspects of their jobs. The demographic make-up of this sample included 58% men with an average age of45.57 years (SD=7.56) and an average organizational tenure of 11.04 years (SD=4.58). In addition to gathering data from theemployees, we also obtained responses from each of their supervisors. Out of a possible 54 supervisors, we received responsesfrom 49 (91% response rate).

We collected data over a three week period. One week prior to the start of the data collection, the agency's Director sent anemail to all members of the state agency introducing the study and requesting their participation. Following this email, theresearchers sent a separate email explaining the study's goals, their Human Subjects rights, and a hotlink to a website hosting thesurvey. Respondents were asked to follow the link and complete the survey before the end of the three week window. At the sametime we emailed all employees, we also emailed their supervisors asking them to provide ratings on each of their subordinates.

4.3. Measures

Responses to the items in our survey were recorded on 5-point Likert scales. Unless otherwise noted, the response anchors forthese items were “strongly disagree” (1) to “strongly agree” (5). In designing the survey for the both samples, we were asked byboth the university alumni organization and the state agency to keep the survey as easy and as simple to complete as possible. Afterreceiving feedback from both organizations, we decided to use the 5-point Likert scales for all scales.

4.4. Subordinate measures

4.4.1. Leader–member exchangeIn sample 1, LMX relationship quality was measured via subordinate responses to Scandura, Graen, and Novak's (1986) seven-

item measure (α=.94). However, we must note that the 5-point Likert scale that was approved by the university alumniorganization and used thereafter was different from the one used by Scandura et al. (1986). A sample item was “My supervisorwould be personally inclined to use his or her power to help me solve problems in my work.” In sample 2, we used Liden andMaslyn's (1998) 12-item (α=.92) multidimensional LMX scale to measure exchange quality. A sample item included “Mysupervisor would defend me to others in the organization if I made an honest mistake.”

4.4.2. EmpowermentIn both samples, employees responded to an adapted version of Spreitzer's (1995) 12-item empowerment scale (α=.87, .87).

This scale was adapted by using a 5-point Likert scale instead of the original 7-point scale used by Spreitzer (1995). A sample itemis “The work I do is meaningful to me.”

4.4.3. Job satisfactionIn sample 1 we measured job satisfaction using Cammann, Fichman, Jenkins, & Klesh's (1979) 3-item scale (α=.89). A sample

item was “All in all, I am satisfied with my job.”

Table 1Confirmatory factor analysis results for samples 1 and 2.

Sample 1 Sample 2

X2 df Xdiff2 CFI NFI RMSEA X2 df Xdiff

2 CFI NFI RMSEA

1-factor 5056 275 .62 .60 .226 3834 434 .63 .59 .2264-factor 2383 269 2673 ⁎⁎⁎ .81 .78 .179 2040 428 1794 ⁎⁎⁎ .80 .80 .1564-factor with 2nd order empowerment 88 50 .98 .97 .0562-factor DVs 28 13 .99 .97 .0872-factor IVs 1788 251 .76 .73 .1992-factor IVs with 2nd order empowerment and LMX 412 243 .97 .93 .067

Note: DV = dependent variable, IV = independent variables, LMX = leader–member exchange.Sample 1 N=244, sample 2 N=158.⁎⁎⁎ pb .001.

375K.J. Harris et al. / The Leadership Quarterly 20 (2009) 371–382

4.4.4. Turnover intentionsIn sample 1, we measured turnover intentions using Seashore, Lawler, Mirvis, & Cammann's (1982) three-item (α=.92) scale.

A sample item was “I will probably look for a new job in the next year.”

4.5. Supervisor measures

4.5.1. Job performanceIn sample 2, Podsakoff and MacKenzie's (1989) four-item (α=.71) scale was used by supervisors to provide ratings of their

subordinates' job performance. A sample item included “This subordinate fulfills all responsibilities required by his/her job.”

4.5.2. Organizational citizenship behaviorsIn sample 2 we used Liden, Wayne, Jaworski, and Bennett's (2004) 3-item (α=.81) OCB scale to have supervisors rate their

subordinates' citizenship behaviors. Supervisors responded to a 5-point scale with anchors of 1 (not at all characteristic) to 5 (verycharacteristic). A sample item was “This subordinate volunteers to do things not formally required by the job.”

4.6. Control variables

In both samples we controlled for organizational tenure (measured in years), gender (females = 1, males = 2), and age in ouranalyses. Previous research has suggested or shown these variables to be related to our outcome variables (e.g., Hom & Griffeth,1995; Spector, 1997; Sturman, 2003).

4.7. Analysis

We tested our first two hypotheses using hierarchical moderated regression analyses. However, the outcome variables insample 2 were rated by supervisors, who each provided ratings on multiple subordinates. Thus, the data in sample 2 were non-independent. As a result, we used hierarchical linear modeling (HLM; Raudenbush, Bryk, Cheong, & Congdon, 2004) to analyze thethird and fourth hypotheses. Both of these analytical procedures included four steps. In the first step, we entered the three controlvariables. In the second step, we entered the mean centered LMX term and in the third step entered the mean centered

Table 2Means, standard deviations, and intercorrelations.

Variable Sample 1Mean

Sample 1SD

Sample 2Mean

Sample 2SD

1 2 3 4 5 6 7 8 9

1. Leader–member exchange 3.61 .88 4.00 .62 – .48 ⁎⁎ .08 .12 − .03 .15 .122. Empowerment 4.13 .56 3.93 .53 .36 ⁎⁎ – .14 .15 − .02 .11 − .053. Organizational tenure 12.82 9.97 11.04 4.58 − .02 .01 – .09 .34 ⁎⁎ .06 − .054. Sex 1.51 .50 1.58 .50 .04 .14 ⁎ .24 ⁎⁎ – .05 − .03 .015. Age 46.59 10.71 45.57 7.56 − .03 .17 ⁎⁎ .36 ⁎⁎ .15 ⁎⁎ –

6. Job satisfaction 4.20 .74 .40 ⁎⁎ .58 ⁎⁎ .01 − .02 .07 –

7. Turnover intentions 2.21 1.25 − .28 ⁎⁎ − .32 ⁎⁎ − .26 ⁎⁎ − .03 − .16 ⁎⁎ − .56 ⁎⁎ –

8. Job performance 4.43 .61 – .59 ⁎⁎9. OCBs 4.24 .75 –

Sample 1 correlations are below the diagonal and sample 2 correlations are above the diagonal.Note: OCBs = organizational citizenship behaviors.Sample 1 N=244, sample 2 N=158.⁎ pb .05.⁎⁎ pb .01.

Table 3Hierarchical moderated regression analyses.

DV = job satisfaction DV = turnover intentions

Step 1 Step 2 Step 3 Step 4 Step 1 Step 2 Step 3 Step 4

Step 1Organizational tenure − .02 − .01 .03 .03 − .24 ⁎⁎ − .24 ⁎⁎ − .26 ⁎⁎ − .26 ⁎⁎Sex − .02 − .04 − .11 ⁎ − .11 ⁎ .04 .05 .08 .09Age .08 .09 − .01 .01 − .08 − .08 − .04 − .05

Step 2Leader–member exchange .40 ⁎⁎ .22 ⁎⁎ .21 ⁎⁎ − .28 ⁎⁎ − .19 ⁎⁎ − .19 ⁎⁎

Step 3Empowerment .52 ⁎⁎ .51 ⁎⁎ − .26 ⁎⁎ − .25 ⁎⁎

Step 4Leader–member exchange ⁎ − .13 ⁎⁎ .10 ⁎Empowerment

Simple slopes t-testsHigh empowerment 4.27 ⁎⁎ −2.99 ⁎⁎Low empowerment .34 −0.48

Total R2 .01 .17 .39 .41 .07 .15 .21 .22Change in R2 .01 .16 ⁎⁎ .22 ⁎⁎ .02 ⁎⁎ .07 ⁎⁎ .08 ⁎⁎ .06 ⁎⁎ .01 ⁎

Note: Standardized coefficients are provided.N=244.⁎ pb .05.⁎⁎ pb .0.

376 K.J. Harris et al. / The Leadership Quarterly 20 (2009) 371–382

empowerment term (Aiken & West, 1991). In the final step, we entered the interaction term formed by multiplying the meancentered LMX and empowerment terms.

5. Results

Prior to testing our hypotheses, we conducted confirmatory factor analyses (CFAs) on both samples in an effort to establish thediscriminant validity of our measures. Establishing discriminant validity was especially important for sample 1 given that all of thedata came from the same source. For sample 2, our goal was to demonstrate that the IVs, which came from the same source,showed discriminant validity from one another and that the DVs, which came from the same source, also were distinct from oneanother. To accomplish these goals, we estimated multiple models with each dataset using LISREL 8.80. For sample 1, we beganwith a 1-factor model where all of the items for the four scales (LMX, empowerment, job satisfaction, and intent to turnover) allloaded on the same factor. Results for this analysis, shown in Table 1, indicate that themodel did not fit the data. Next, we estimateda 4-factor model, with one factor representing each of our variables. While model fit (see Table 1) improved, the model producedonly moderate fit. Finally, we ran a model in which the empowerment scale was modeled as a second-order factor composed of 4subscales (meaning, competence, autonomy, and impact). As shown in Table 1, this model fit the data well. As a whole, theseresults suggest that the scales used in sample 1 were distinct from one another and that collapsing the empowerment items intoone scale was appropriate.

We estimated a similar series of models for sample 2. We began with a 1-factor model (see Table 1) which did not fit the datawell. We followed this with a 4-factor model in which each variable was represented by a different factor: LMX, empowerment,OCB, and job performance. Again, while model fit improved, the overall fit was only moderate. These results were not surprising

Fig. 1. Interaction between LMX and empowerment predicting job satisfaction in sample 1.

Fig. 2. Interaction between LMX and empowerment predicting turnover intentions in sample 1.

377K.J. Harris et al. / The Leadership Quarterly 20 (2009) 371–382

given that the data for the IVs and DVs came from different sources. Thus, we next exploredwhether the variables from each sourcewere distinguishable from one another. To do this we ran a 2-factor model for the independent variables and a 2-factor model forthe dependent variables. As shown in Table 1, the fit statistics for the dependent variable model were acceptable, but the model forthe IVs indicated improvements could be made. Following our logic for sample 1, we reran the IV model as a second-order modelthat modeled both empowerment and LMX as higher order factors each composed of 4 subscales. The fit for this model (seeTable 1) was acceptable. In total, these analyses suggest that the scales provided by the same raters were distinct from one anotherand that using the empowerment and LMX scales unidimensionally was appropriate.

Table 2 provides themeans, standard deviations, and intercorrelations among variablesmeasured for both samples in our study. Afew correlations are of specific interest. The first is the significant association between LMX and empowerment (r=.36 for sample 1,r=.48 for sample 2). These correlations show that thesevariables are related, but that there is considerable uniqueness between them.Other correlations of interest are between LMX and the outcomes examined in this study. Table 2 shows that LMX was significantlyrelated to job satisfaction (r=.40) and turnover intentions (r=− .28), but not to job performance (r=.15) and OCBs (r=.15). Theseresults, especially for the performance-related variables, make the examination of moderators all the more important.

Table 3 provides the results of our hierarchical moderated regression analyses. Step 2 in each analysis shows that LMX waspositively and significantly related to job satisfaction (β=.40, pb .01, ΔR2=.16) and negatively and significantly related toturnover intentions (β=− .28, pb .01, ΔR2=.08). In step 3 of each analysis, empowerment was positively and significantly relatedto job satisfaction (β=.52, pb .01, ΔR2=.22) and negatively and significantly related to turnover intentions (β=− .26, pb .01,ΔR2=.06). Finally, the interaction terms were entered in step 4 in each analysis. As can be seen, the LMX⁎empowermentinteraction was significantly related to both job satisfaction (β=− .13, pb .01, ΔR2=.02) and turnover intentions (β=.10, pb .05,

Table 4Hierarchical linear modeling results.

DV = job performance DV = OCBs

Step 1 Step 2 Step 3 Step 4 Step 1 Step 2 Step 3 Step 4

Step 1Organizational tenure .00 .00 .00 − .00 .00 − .00 − .00 − .00Sex − .04 − .06 − .07 − .06 .02 − .00 .01 .03Age .00 .00 .00 .00 .00 .00 .00 .01

Step 2Leader–member exchange .20 ⁎ .16 t .15 t .18 t .22 t .21 t

Step 3Empowerment .11 .15 − .11 − .06

Step 4Leader–member exchange ⁎ − .28 ⁎ − .37 t

EmpowermentSimple slopes t-testsHigh empowerment 2.49 ⁎⁎ 2.18 ⁎Low empowerment −0.18 −0.12

Total R2 .01 .06 .06 .08 .00 .02 .02 .04Change in R2 .01 .05 ⁎ .00 .02 ⁎ .00 .02 t .00 .02 t

Note: Standardized coefficients are provided. OCBs = organizational citizenship behaviors.N=158.

t pb .10.⁎ pb .05.⁎⁎ pb .0.

Fig. 3. Interaction between LMX and empowerment predicting job performance in sample 2.

378 K.J. Harris et al. / The Leadership Quarterly 20 (2009) 371–382

ΔR2=.01). However, in order to determine support for Hypotheses 1 and 2, we graphed our interactions. Fig.1 provides a graphicalrepresentation of the interaction with job satisfaction as the dependent variable. As seen in this figure, the greatest rate of changefor the job satisfaction line slope occurred for those respondents who reported low empowerment. Conversely, there was littlechange in the job satisfaction line slope for those respondents who reported high empowerment. Additionally, the positiverelationship between LMX and job satisfaction was strongest when empowerment was lowest. Results from simple slopes tests,shown at the bottom of Table 3, confirm this interpretation. Taken together, these results provide support for Hypothesis 1.

Fig. 2 provides a graphical depiction of the LMX⁎empowerment interactionwith turnover intentions as the outcome. As seen inthis figure, the rate of change for the turnover intentions line slope is strongest for those respondents who reported lowempowerment, while the rate of change for the turnover intentions line slope for high empowerment was less strong. Interpretingthis interaction, we again see the moderating influence of empowerment on the LMX quality–outcomes relationship. In this case,the respondents reporting low empowerment coupled with low quality LMX relationships reported the highest turnoverintentions. The result from our simple slopes test on this interaction, which is at the bottom of Table 3, supports this explanation.These results support Hypothesis 2.

Table 4 depicts ourHLM results testing Hypotheses 3 and 4. Step 2 in each analysis shows that LMXwas positively and significantlyrelated to job performance (γ=.20, pb .05, ΔR2=.05) and OCBs (γ=.18, pb .10, ΔR2=.02). In step 3 of each analysis, empowermentwasnot significantly related to either job performance orOCBs. Finally, the interaction termswere entered in step 4 in each analysis. Ascan be seen, the LMX⁎empowerment interactionwas significantly related to both job performance (γ=− .28, pb .05, ΔR2=.02) andOCBs (γ=− .37, pb .10, ΔR2=.02). However, in order to determine support for Hypotheses 3 and 4, we again graphed the patterns ofour interactions. Fig. 3 provides a graphical representation of the interactionwith job performance as the dependent variable. As seenin this figure, a crossover interaction occurred whereby the job performance line slope for respondents reporting low empowermentpositively increases as the quality of LMX relationships increases. Interestingly, the job performance line slope for respondentsreporting high empowerment is virtually unchanged as the quality of the LMX relationship increases. That is, the positive LMX–jobperformance relationship is strongest when empowerment is low. At the bottom of Table 4 can be seen the results from our simpleslopes tests which confirms this interpretation. These results provide support for Hypothesis 3.

Fig. 4. Interaction between LMX and empowerment predicting OCBs in sample 2.

379K.J. Harris et al. / The Leadership Quarterly 20 (2009) 371–382

Finally, Fig. 4 illustrates the interaction with OCBs as the outcome. This graph depicts another crossover interaction. The OCBline slope for those respondents reporting low empowerment rapidly and positively increases as the quality of the LMXrelationship increases. Moreover, the OCB line slope for respondents reporting high empowerment is almost unchanged as LMXquality increases. This finding is similar to the pattern of results found for Hypothesis 3, as the positive LMX–OCB relationship wasstrongest when respondents reported lower levels of empowerment. Results from simple slopes tests, shown at the bottom ofTable 4, confirm this understanding. These results provide support for Hypothesis 4.

6. Discussion

The current study set out to examine themoderating effect of empowerment on the relationships between LMX quality and thejob outcomes of job satisfaction, turnover intentions, job performance, and OCBs. Overall, the results provided support for themoderating impact of empowerment. Specifically, we found that LMX relationship quality mattered the most for employees whofelt little empowerment. Both LMX theory and JCT theory posit that LMX quality and empowerment, respectively, should beassociated with more positive outcomes (higher job satisfaction, job performance, and OCBS, or lower turnover intentions), andprevious results have established these linkages (e.g., Aryee & Chen, 2006; Gerstner & Day, 1997; Graen & Uhl-Bien, 1995;Schriesheim et al., 1999; Spreitzer et al., 1997). The interactive effects found in the present study showa slight synergistic effect, butmore importantly, that when empowerment (the moderator) is lower, the relationships between LMX and desired outcomes arestronger.

These findings can be explained by elements of JCT and social exchange theory. As expected, when both LMX quality andempowerment were low the most negative outcomes resulted, and in general, when both variables were high the most positiveoutcomes resulted. More interestingly, our results showed that when employees lack motivation in their jobs (lowempowerment), the LMX relationship becomes more important. Essentially, if an employee is high in empowerment, the LMXrelationship becomes of less importance because regardless of the relationship between an employee and his/her supervisor, thejob provides the motivation which is associated with more positive outcomes.

However, when empowerment is low, employees look to other workplace aspects to receive the benefits they fail to receivefrom empowerment. One place that employees are likely to look at is their relationship with their supervisors. This logic is in linewith our findings and can be exchanged by social exchange theory, which suggests that when empowerment is low, socialexchanges may be of heightened importance. In our study, we found that when empowerment was lacking, the LMX relationshipbecame more vital. These findings are likely due to the fact that supervisors are able to offer benefits that can keep employeesmotivated in a different way thanworkplace empowerment (Liden et al., 1997). However, it is worth noting that the highest levelsof OCBs resulted when LMX quality was high and empowerment was low. This finding was surprising as we expected OCBs to behighest when both empowerment and LMX relationship quality were high (Ilies et al., 2007). Thus, we invite future researchers toexamine this relationship and better determine the reasons for our unexpected finding.

While the findings from this study make a unique contribution to both the LMX and empowerment research streams, theprimary contribution of our research is the joint effect of these variables, as empowerment moderates the LMX–outcomeassociations. In particular, we used two samples and examined four outcomes, two of which were provided by self-reports and twoprovided by supervisor-reports, and found that empowerment exhibited a similar moderating effect on the relationships betweenLMX and the consequences we examined.

These investigations have been called for by previous researchers (e.g., Liden et al., 2000; Schriesheim et al., 1999), and help toclarify previous findings that have failed to find interactions between LMX and aspects of the job itself (usually job characteristics)(e.g., Graen, Novak, & Sommerkamp,1982; Seers & Graen,1984). For example, whereas Seers and Graen (1984) focused specificallyon the interaction between job characteristics and LMX and found weak empirical support, the present study examined thebroader conceptualization of empowerment. Graen et al. (1982) found similar results examining the interaction between jobdesign training and LMX, finding only main effects of LMX on satisfaction and performance. These previous null findings relating tothe interaction of LMX and job characteristics or job design training are tangentially related to the present study in thatempowerment is theoretically rooted in JCT; however, the construct of empowerment differs from empirical applications of jobcharacteristic models. Empowerment focuses on how employees think about their jobs, effectively linking the criticalpsychological states outlined in JCT. That is, empowerment does not focus on the components of the job, as does JCT, as muchas it focuses on the affective outcomes related to the job. Thus, we feel that our results contribute to the literature by showing thatempowerment moderates the relationships between LMX and four outcomes, such that the relationships become stronger whenempowerment is lower. Additionally, we have increased confidence in our results as we found them in two different samples, withoutcomes rated by both the focal employees themselves and their supervisors.

6.1. Practical implications

The results of this study have practical implications for managers. First, our findings point to the fact that managers should notjust train supervisors (Graen, 1989) to increase the quality of their LMX relationships with subordinates or implementempowerment activities (Conger & Kanungo, 1988), but they need to do both but for different reasons. In terms of affectivevariables (job satisfaction and turnover intentions), we observed the most positive/least negative levels of these outcomes whenLMX relationship quality and empowermentwere high. However, as long as empowerment was high, LMX relationship quality hadrelatively minimal impact. Thus, the practical focus for managers should be on empowerment. But, in cases where employees are

380 K.J. Harris et al. / The Leadership Quarterly 20 (2009) 371–382

not given opportunities to feel highly empowered, high quality LMX relationships can help compensate for some of the drawbacksof low empowerment.

Similarly, in terms of employee task and contextual performance, our results suggest that managers should focus on increasingemployee perceptions of empowerment. However, if this is not possible, high quality LMX relationships seem to provide benefits andthe resulting high levels of performance that are not provided throughempowerment. Thus,when empowerment is lacking, it appearsas though training programs targeting the development of LMX relationships (Graen, 1989) should be associated with increases inemployee performance levels. In total, these findings indicate that empowerment is important and in most cases it is preferable formanagers to focus on increasing empowerment levels, even if this is at the expense of promoting higher quality LMX exchanges.However,when empowerment is lacking, the LMX relationship becomes of additional importance. Thus,managers need to be aware oflow empowerment situations, and in these cases do everything in their power to help develop high quality exchanges.

6.2. Limitations of the study

Although this study makes a number of contributions to the extant literature, there are limitations that need to beacknowledged. One is the fact that our results may have problems related to common source and common method variance(CMV). In particular, all of the variables in sample 1 were examined from the same source. We attempted to minimize theseconcerns by examining sample 2, which had two supervisor-rated variables as outcomes. Additionally, we conducted a Harmonone-factor test (Podsakoff & Organ, 1986) on the data in sample 1, and the results of this test, which found that the first (primary)factor explained only 34% of the variance, provided evidence that CMVwas not a pervasive problem in sample 1.We also conductedan extensive series of confirmatory factor analyses on the data collected from both samples that established the discriminantvalidity of our measures. Finally, extant research has shown that CMV actually decreases the probability of finding significantinteractions (Wall, Jackson, Mullarkey, & Parker, 1996).

A second limitation is that our analyses showed that the interaction terms explained between 1 and 2% of the variancerespectively. Although relatively small, these results are in line with previous research on interactions in non-experimental studies(Champoux & Peters, 1987). Additionally, previous research has shown that the practical impact of small effect sizes can besubstantial (Abelson, 1985; Fichman, 1999).

A third limitation is that we were unable to examine between group vs. within group effects (Dansereau & Yammarino, 2000).In our first sample, each respondent reported to a separate supervisor. In the second sample, the employees did not work in groups,but worked separately, and then reported to the same supervisors. For these reasons, it was either impossible or inappropriate tomeasure these effects. However, we hope future researchers will examine this question (Boies & Howell, 2006) as it will provideinteresting insights into the moderating effect of empowerment on LMX–outcome relationships.

A final limitation is that our study does not test for the causality of our proposed relationships. Our theoretical argumentsposition LMX as the independent variable and empowerment as the moderator. However, it is possible that if someone likes his/her job more or less (job satisfaction), is thinking about leaving or not leaving the organization (intent to turnover), or is a betterperformer, this could lead to higher or lower levels of LMX and/or empowerment. With this being said, we feel that our study,which reports results consistent with other studies on these variables (Schriesheim et al., 1999; Spreitzer et al., 1997), and ourtheoretical arguments (e.g., Cropanzano & Mitchell, 2005; Hackman & Oldham, 1976), which better position LMX andempowerment as predictors, minimize this limitation.

6.3. Directions for future research

The results of our study lead to a number of fruitful directions for future research. First, we invite future researchers to examineour hypotheses in a longitudinal study (e.g., Bauer, Erdogan, & Liden, 2006). This would help to answer the questions related tohow LMX relationships and empowerment programs change over time and how the moderating effect of empowerment becomeseithermore or less pronounced. Another direction is to examine these phenomena at different levels (Yammarino, Dionne, Chun, &Dansereau, 2005). The respondents in our first sample were all from different organizations and reported to different supervisors,thus preventing us from employing any multilevel analyses. However, we did examine our second sample using a multilevelanalytical technique. Unfortunately though, we were unable to access multiple different levels in the organization.

With that being said it would be interesting to examine how a supervisor's LMX relationship with his/her boss plays a role inthe subordinate's LMX relationship (Tangirala, Green, & Ramanujam, 2007). We suspect that subordinates make assumptions oftheir supervisors' relationship qualities (i.e., are they liked, disliked) and based on these assumptions, they act in different ways.Additionally, LMX may be more or less impacted by empowerment based on characteristics of the supervisor or even thesupervisor's supervisor. In particular, we invite future researchers to examine supervisor's or supervisor's supervisor's power,influence, and political skill (e.g., Ferris, Treadway, Perrewé, Brouer, Douglas, & Lux, 2007; Harris, Kacmar, Zivnuska, & Shaw, 2007;Sparrowe & Liden, 2005) to determine their impact on the relationships between LMX, empowerment, and job outcomes.

Similar level issues have been shown in empowerment research. In particular, Seibert et al. (2004) showed the utility inexamining how an overall empowerment culture at the organizational level can have a significant impact on empowermentperceptions at the individual level. Finally, we believe that research on other outcome variables is warranted. In this study, weinvestigated four key organizational outcomes, including a job attitude, a behavioral intention, and two behaviors. Although theseoutcomes are important and cover a range of consequences, it would be insightful to determine if empowerment similarlymoderates the relationships between LMX and creativity, actual turnover, and stress–strain related reactions.

381K.J. Harris et al. / The Leadership Quarterly 20 (2009) 371–382

7. Conclusion

Our investigation found that for four different job outcomes, including an attitude, behavioral intention, and behaviors,empowerment moderated the relationships between LMX and these consequences. In particular, our findings revealed that LMXbecomes especially important when empowerment is lacking. This study extends previous research by examining the joint impactof LMX relationship quality and empowerment, which had not been previously investigated, and showed that this moderatingeffect was present in two samples with outcomes that were self and supervisor-rated. We feel that these results are interesting andwe hope they will encourage future researchers to continue to investigate how these variables work together.

References

Abelson, R. (1985). A variance explanation paradox: When a little explains a lot. Psychological Bulletin, 97, 129−133.Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting interactions. Newbury Park, CA: Sage.Aryee, S., & Chen, Z. X. (2006). Leader–member exchange in a Chinese context: Antecedents, the mediating role of psychological empowerment and outcomes.

Journal of Business Research, 59, 793−801.Bauer, T. N., Erdogan, B., & Liden, R. C. (2006). A longitudinal study of the moderating role of extraversion: Leader–member exchange, performance, and turnover

during new executive development. Journal of Applied Psychology, 91, 298−310.Blau, P. M. (1964). Exchange and power in social life. New York: Wiley.Boies, K., & Howell, J. M. (2006). Leader–member exchange in teams: An examination of the interaction between relationship differentiation and mean LMX in

explaining team-level outcomes. Leadership Quarterly, 17, 246−257.Cammann, C., Fichman, M., Jenkins, D., & Klesh, J. (1979). The Michigan organizational assessment questionnaire. Ann Arbor, MI: University of Michigan.Champoux, J. (1991). A multivariate test of the job characteristics theory of work motivation. Journal of Organizational Behavior, 12, 431−446.Champoux, J., & Peters, W. (1987). Form, effect size, and power in moderated regression analysis. Journal of Occupational Psychology, 60, 243−255.Conger, J. A., & Kanungo, R. N. (1988). The empowerment process: Integrating theory and practice. Academy of Management Review, 13, 471−482.Cropanzano, R., & Mitchell, M. S. (2005). Social exchange theory: An interdisciplinary review. Journal of Management, 31, 874−900.Dansereau, F., & Yammarino, F. J. (2000). Within and between analysis: The varient paradigm as an underlying approach to theory building and testing. In K. J. Klein

& S. W. J. Kozlowski (Eds.), Multilevel theory, research, and methods in organizations: Foundations, extensions, and new directions (pp. 425−466). San Francisco,CA: Jossey-Bass.

Dienesch, R. M., & Liden, R. C. (1986). Leader–member exchange model of leadership: A critique and further development. Academy of Management Review, 11,618−634.

Eby, L. T., Butts, M., & Lockwood, A. (2003). Predictors of success in the era of the boundaryless career. Journal of Organizational Behavior, 24, 689−708.Elicker, J. D., Levy, P. E., & Hall, R. J. (2006). The role of leader–member exchange in the performance appraisal process. Journal of Management, 32, 531−551.Epitropaki, O., & Martin, R. (2005). From ideal to real: A longitudinal study of the role of implicit leadership theories on leader–member exchanges and employee

outcomes. Journal of Applied Psychology, 90, 659−676.Ferris, G. R., Treadway, D. C., Perrewé, P. L., Brouer, R. L., Douglas, C., & Lux, S. (2007). Political skill in organizations. Journal of Management, 33, 290−320.Fichman, M. (1999). Variance explained: Why size does not (always) matter. Research in Organizational Behavior, 21, 295−331.Gerstner, C. R., &Day, D. V. (1997).Meta-analytic reviewof leader–member exchange theory: Correlates and construct issues. Journal of Applied Psychology,82, 827−844.Graen, G. B. (1989). Unwritten rules for your career. New York: Wiley.Graen, G. B. (2004). New frontiers of leadership, LMX leadership: The series. Greenwich, CT: Information Age Publishing.Graen, G. B., Novak, M., & Sommerkamp, P. (1982). The effects of leader–member exchange and job design on productivity and satisfaction: Testing a dual

attachment model. Organizational Behavior and Human Performance, 30, 109−131.Graen, G. B., & Uhl-Bien, M. (1995). Development of leader–member exchange (LMX) theory of leadership over 25 years: Applying a multi-level multi-domain

perspective. Leadership Quarterly, 6, 219−247.Hackman, J. R., & Oldham, G. R. (1976). Motivation through the design of work: Test of a theory. Organizational Behavior and Human Performance, 16, 250−279.Harris, K. J., Kacmar, K. M., Zivnuska, S., & Shaw, J. D. (2007). The impact of political skill on impression management effectiveness. Journal of Applied Psychology, 92,

278−285.Hom, P. W., & Griffeth, R. W. (1995). Employee turnover. Cincinnati: South/Western.Homans, G. C. (1958). Social behavior as exchange. American Journal of Sociology, 63, 597−606.Ilies, R., Nahrgang, J., & Morgeson, F. (2007). Leader–member exchange and citizenship behaviors: A meta-analysis. Journal of Applied Psychology, 92, 269−277.Kahn, R. L., Wolfe, D. M., Quinn, R. P., Snoek, J. D., & Rosenthal, R. A. (1964). Organizational stress: Studies in role conflict and ambiguity. New York: Wiley.Liden, R. C., Erdogan, B., Wayne, S. J., & Sparrowe, R. T. (2006). Leader–member exchange, differentiation, and task interdependence: Implications for individual and

group performance. Journal of Organizational Behavior, 27, 723−746.Liden, R. C., & Maslyn, J. M. (1998). Multidimensionality of leader–member exchange: An empirical assessment through scale development. Journal of Management,

24, 43−72.Liden, R. C., Sparrowe, R. T., &Wayne, S. J. (1997). Leader–member exchange theory: The past and potential for the future. In G. R. Ferris (Ed.), Research in personnel

and human resource management, vol. 15 (pp. 47—120). Greenwich, CT: JAI Press.Liden, R. C., Wayne, S. J., Jaworski, R. A., & Bennett, N. (2004). Social loafing: A field investigation. Journal of Management, 30, 285−304.Liden, R. C., Wayne, S. J., & Sparrowe, R. T. (2000). An examination of the mediating role of psychological empowerment on the relations between the job,

interpersonal relationships, and work outcomes. Journal of Applied Psychology, 85, 407−416.Likert, R. (1961). New patterns of management. New York: McGraw-Hill.Manzoni, J., & Barsoux, J. (2002). The set-up-to-fail syndrome. How good managers cause great people to fail. Boston: Harvard Business School Press.Martin, R., Thomas, G., Charles, K., Epitropaki, O., & McNamara, R. (2005). The role of leader–member exchanges in mediating the relationship between locus of

control and work reactions. Journal of Occupational and Organizational Psychology, 78, 141−147.Podsakoff, P. M., & MacKenzie, S. B. (1989). A second generation measure of organization citizenship behavior. Working paper Bloomington: Indiana University.Podsakoff, P. M., & Organ, D. W. (1986). Self-reports in organizational research: Problems and prospects. Journal of Management, 12, 531−544.Raudenbush, S., Bryk, A., Cheong, Y. F., & Congdon, R. (2004). HLM6: Hierarchical linear and nonlinear modeling. Lincolnwood, IL: Scientific Software International.Scandura, T., Graen, G., & Novak, M. A. (1986). When managers decide not to decide autocratically: An investigation of leader–member exchange and decision

influence. Journal of Applied Psychology, 71, 579−584.Schriesheim, C. A., Castro, S. L., & Cogliser, C. C. (1999). Leader–member exchange (LMX) research: A comprehensive review of theory, measurement, and data-

analytic practices. Leadership Quarterly, 10, 63−113.Seashore, S. E., Lawler, E. E., Mirvis, P., & Cammann, C. (1982). Observing and measuring organizational change: A guide to field practice. New York: Wiley.Seers, A., & Graen, G. B. (1984). The dual attachment concept: A longitudinal investigation of the combination of task characteristics and leader–member exchange.

Organizational Behavior and Human Performance, 33, 283−306.Seibert, S. E., Crant, J. M., & Kraimer, M. L. (1999). Proactive personality and career success. Journal of Applied Psychology, 84, 416−427.Seibert, S. E., Silver, S. R., & Randolph,W. A. (2004). Taking empowerment to the next level: Amultiple-level model of empowerment, performance, and satisfaction.

Academy of Management Journal, 47, 332−349.

382 K.J. Harris et al. / The Leadership Quarterly 20 (2009) 371–382

Sparrowe, R. T., & Liden, R. C. (2005). Two routes to influence: Integrating leader–member exchange and network perspectives. Administrative Science Quarterly, 50,505−535.

Spector, P. E. (1997). Job satisfaction: Application, assessment, causes, and consequences. Thousand Oaks, CA: Sage.Spreitzer, G. M. (1995). Psychological empowerment in the workplace: Dimensions, measurement, and validation. Academy of Management Journal, 38,

1442−1465.Spreitzer, G. M., Kizilos, M. A., & Nason, S. W. (1997). A dimensional analysis of the relationship between psychological empowerment and effectiveness,

satisfaction, and strain. Journal of Management, 23, 679−704.Sturman, M. C. (2003). Searching for the inverted U-shaped relationship between time and performance: Meta-analyses of the experience/performance, tenure/

performance, and age/performance relationships. Journal of Management, 29, 609−640.Tangirala, S., Green, S., & Ramanujam, R. (2007). In the shadow of the boss's boss: Effects of supervisors' upward exchange relationships on employees. Journal of

Applied Psychology, 92, 309−320.Tierney, P., Farmer, S., & Graen, G. (1999). An examination of leadership and employee creativity: The relevance of traits and relationships. Personnel Psychology, 52,

591−620.Wall, T. D., Jackson, P. R., Mullarkey, S., & Parker, S. K. (1996). The DC-model of job strain: A more specific test. Journal of Occupational and Organizational Psychology,

69, 153−166.Yammarino, F. J., Dionne, S. D., Chun, J. U., & Dansereau, F. (2005). Leadership and levels of analysis: A state-of-the-science review. Leadership Quarterly, 16,

879−919.