You are on page 1of 26

ARTICLES

Employees’ Perceived Use of


Leader Power and Implications
for Affect and Work Intentions

Drea Zigarmi, Taylor Peyton Roberts, W. Alan Randolph

The concept of power in organizations has been studied at both the macro
level (analyses of structural systems or policies) and at the micro level
(individual perceptions). In this study, we examine employee perceptions of
their leader’s use of power at the individual/psychological level. Applying
social cognitive theory, employee perceptions of their leader’s use of various
forms of power were explored in relationship to employees’ negative or
positive affect and corresponding work intentions. Structural equation
modeling was used to examine data from 651 employees. Positive and
negative affect mediated employees’ perceptions of their managers’ use of
various power bases and five work intentions: intentions to perform, to
endorse the organization and its leadership, to stay in the organization, to
use discretionary effort, and to be an organizational citizen. Implications
for practice and future research are discussed.

Key Words: employee affect, appraisal model, power, work intentions

Power and politics are an unavoidable reality in today’s organizations (Clegg,


Courpasson, & Phillips, 2011; Pfeffer, 1992). A preponderance of the litera-
ture on power in the past 75 years has been philosophical or sociological
in nature (Elias, 2008). Much of the literature has addressed organizational
power and politics, “analyzing the conditions of embeddedness of organi-
zational politics in the social structure of corporations” (Clegg et al., 2011,
p. 387). These structural perspectives on power (Pfeffer, 1992) are concerned
with the analysis of the total organization, the foundation of authority sys-
tems, and the character of legitimate power relations (e.g., DuBrin, 2009;
Haugaard & Clegg, 2012; Lunenberg, 2012; Pfeffer, 1992, 2011).
As the topic of power has begun to appear in the organizational literature
over the past 50 years, the importance of the concept of social power has
HUMAN RESOURCE DEVELOPMENT QUARTERLY, vol. 26, no. 4, Winter 2015 © 2015 Wiley Periodicals, Inc.
Published online in Wiley Online Library (wileyonlinelibrary.com) • DOI: 10.1002/hrdq.21216 359
360 Zigarmi, Roberts, Randolph

become popular with management scholars and industrial/organizational psy-


chologists (Elias, 2008). Power is now frequently examined empirically in the
psychological and leadership literature according to the types of power com-
monly enacted by leaders, when types of power are effectively used, and why
leaders tend to move away from effective shared leadership in favor of act-
ing in self-interest (e.g., DeCelles, DeRue, Margolis, & Ceranic, 2012; French
& Raven, 1959; Lammers & Stapel, 2009; Maner & Mead, 2010; Pitesa &
Thau, 2013). A second prevalent line of investigation has emerged through
the examination of the effects of the distribution of power on employee per-
ceptions of empowerment (e.g., Randolph & Kemery, 2011; Seibert, Silver, &
Randolph, 2004; Tjosvold, 2006).
An analysis of publications in the four journals connected to the
Academy of Human Resource Development (AHRD) found very few empiri-
cal articles on leadership and power in the last 10 years. First, the Advances in
Developing Human Resources journal—which provides a central focus within
each issue—revealed no single issue on the topic of leadership and power. Yet
when focusing on leadership development and training curriculums within
a single journal issue, authors frequently advocated for increasing leaders’
awareness of the pros and cons resulting from the use of various forms of
power (e.g., Callahan, Whitener, & Sandlin, 2007; Keller, 2007). Second, our
examination of publications in the Human Resource Development Review iden-
tified no article in the past 10 years primarily focused on the subject of lead-
ership and the use of various forms of power as conceptualized by French
and Raven (1959). But it is sometimes recommended by HRD practitioners
that leaders build their skills in the various forms of power (e.g., Gubbins
& Garavan, 2005; Yorks, 2004). Third, an examination of Human Resource
Development Quarterly publications showed only two citations to French and
Raven’s social power model (i.e., Germain & Tejeda, 2012; Kochery, 1993).
Without speaking to French and Raven’s model specifically, Vince (2014)
recently addressed power, emotion, and HRD in a provocative, invited article
featured in the winter issue of the Human Resource Development Quarterly.
Finally, our analysis of published articles in the Human Resource Development
International revealed only three articles that refer to forms of power used by
leaders in the past 15 years (i.e., Byrd, 2008; Hopfl, 1999; Montesino, 2002).
At the present time, the topic of leadership and power does not seem to have
received much attention in the field of HRD, yet the adept use of power is a
key skill that leaders and HRD practitioners should demonstrate to be maxi-
mally effective in organizations.

Purpose of the Study


This article seeks to augment current understanding of structural power
at the organizational level by examining how structural forms of power as
implemented by leaders affect people at the individual psychological level.

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


Employees’ Perceived Use of Leader Power and Implications 361

As HRD strives to maximize the performance, retention, and motivation of


employees, HRD practitioners should consider how the daily use and mis-
use of power by leaders influences such objectives. Because the leader and
follower are ensconced in a shared context in relation to the institution and
its tasks (Krantz, 2006), one purpose of this piece is to reignite awareness of
the importance of considering the place and function leadership power plays
in today’s organizations by empirically examining its potential psychological
impact.
As work context may create complex power relationships, it is para-
mount for HRD and leadership researchers to understand the possible posi-
tive and negative impact that various forms of power may have on employee
outcome variables. Regarding the use of power in organizations, Mossholder,
Bennet, Kemery, and Wesolowski (1998) describe and explain an experience
commonly felt by employees: “ … as supervisors employ various bases of
power, subordinates will form evaluative perceptions regarding the behavior
exhibited by a power holder. These perceptions likely then become a critical
factor in determining subordinates’ subsequent reactions” (p. 534).
We believe this article also addresses the call made by Vince (2014), who
cogently argues that power arises from the interplay between structure and
agency. Vince (2014) highlights that power is present in organizations as a
consequence of social relationships and is implicit within interpersonal rela-
tions, and that power relations and emotions are highly connected. According
to Vince (2014), in order for HRD practitioners to understand the learn-
ing and development capacity of organizational members, it is important to
“engage critically with the ways in which everyday emotions and power rela-
tions undermine as well as support this aim” (p. 418). By examining the rela-
tionship between employees’ perceived use of various forms of leader power
and employees’ affective responses, HRD researchers may gain insight into the
positive or negative impact of various forms of power.
Podsakoff and Schriesheim (1985) called for the examination of the
independent contribution that different types of power make to account for
variance in subordinate outcome variables. By exploring the critical connec-
tion between subordinates’ evaluative perceptions of various kinds of manage-
rial power and their psychological reactions to such power, this study aims to
fill this important gap that remains in the literature.
In particular, this study examines the impact of individuals’ cognitive
perceptions, affect, and intentions that may result from experiencing various
forms of power practiced by their leaders. Through the examination of the
individual’s cognitive and affective appraisal process, we aim to gain insight
into key psychological processes at work and derive recommendations for
HRD practitioners going forward. This article begins by first providing the
theoretical context for the variables examined in this study. Second, an over-
view of the literature on power, affect, and intention sets the basis for our pro-
posed hypotheses. Third, the study’s methodology is presented. Results and

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


362 Zigarmi, Roberts, Randolph

discussion follow, and the article concludes with comments on the limitations
of this work and notable implications for HRD practice.

Theoretical Approach
In line with Aguinis, Nesler, Quigley, and Tedeshi (1994), we propose that
social cognitive theory provides a promising framework through which
employees’ evaluative processes can be investigated within the context of
managerial power. Social cognitive theory holds that human beings are agents
of their own future who adjust and adopt new behaviors in response to their
environment (Bandura, 1986). Because individuals are capable of forethought,
self-regulation, self-reflection, vicarious experiencing, and symbolizing, they
are able to view behaviors as if they are freely chosen according to expecta-
tions about future outcomes and possibilities (Bandura, 1986).
Agency is not only about self-interest, but also about learning and using
social influences to modify behavior in response to what is experienced. As
such, individuals cognitively and affectively evaluate their experiences from
the perspective of their own general welfare by connecting work events to
their emotions, values, and desired outcomes (e.g., Bagozzi, 1992; Lazarus,
1991), as well as to their future intentions and behaviors.
Various Forms of Power as Cognitive Perception
Cognition is concerned with the acquisition of mental schema, or patterns of
thought that pertain to feelings, images, ideas, and descriptions connected
with the appraisal of work experiences (L. A. James & James, 1989; Lord
& Kernan, 1987). In particular, work cognition comprises the development
of mental descriptions and valuations through repeated appraisal processing
of work events (L. A. James & James, 1989; Lord & Emrich, 2001; Wofford
& Goodwin, 1990). Work cognitions, or schemata, are mental structures,
hierarchically formed through connections between perceived schema of
present experience and the remembrance of past experience (L. A. James &
James, 1989; Lord & Emrich, 2001; Wofford & Goodwin, 1990). A number
of researchers in the field of management and organizational literature have
identified various forms of social power as critical cognitive structures (e.g.,
Aguinis et al., 1994; Farmer & Aguinis, 2005; French & Raven, 1959).
Since the basis of social power was initially published by French and
Raven (1959), there have been number of studies investigating the use and
perceptions of the five forms of power (i.e., reward, legitimate, coercive, refer-
ent, and expertise power) at the individual, psychological level (e.g., Farmer
& Aguinis, 2005). In the past 50 years, the conceptual approach to studying
the different types of power has largely remained in keeping with French and
Raven’s initial fivefold taxonomy (Elias, 2008). While there have been con-
troversial aspects of this line of thinking such as single-item measurement,
response bias potential, and concept overlap (cf. Podsakoff & Schriesheim,

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


Employees’ Perceived Use of Leader Power and Implications 363

1985), the fivefold taxonomy advocated by French and Raven continues to be


the predominant framework in the literature for the analysis of power (Elias,
2008; Podsakoff & Schriesheim, 1985).
French and Raven (1959) delineated five different forms of social power:
coercive power, reward power, legitimate power, expert power, and refer-
ent power. Coercive power is defined as the capacity of the leader to pun-
ish the follower if there is a failure to conform to desired outcomes (Raven,
Schwarzwald, & Koslowsky, 1998). The strength of coercive power corre-
sponds to the degree of the punishment possible, and its use often results in a
negative relationship between leader and follower (Podsakoff & Schriesheim,
1985). Reward power stems from a promise of monetary or nonmonetary
compensation (Raven et al., 1998). The intensity of reward power increases
with the magnitude of the rewards possible as well as its appeal to the
receiver (Podsakoff & Schriesheim, 1985). Legitimate power originates from
the subordinate’s perceptions of the leader’s right to influence (Raven et al.,
1998). The strength of legitimate power is attributable to the internalized
values the follower has concerning the authority or right of the influencer
to be the influencer (Podsakoff & Schriesheim, 1985). Expert power relies
on perceptions the follower holds regarding the influencer’s superior knowl-
edge (Raven et al., 1998). The degree of this power depends on the extent
of knowledge or expertise the follower perceives the influencer to have in a
given area (Podsakoff & Schriesheim, 1985). Finally, referent power is based
on the follower’s identification with the influencer (Raven et al., 1998). The
strength of this power is based on the degree to which the subordinate’s own
self-identity is enhanced through interaction with the influencer or the desire
of the follower to be similar to, or associated with, the influencer (Podsakoff
& Schriesheim, 1985).
Over the years, existing literature demonstrates how the use of these vari-
ous forms of power at the individual level relates to outcomes. Specifically, soft
forms of power (e.g., referent power, expert power) are connected to higher
levels of organizational commitment, job satisfaction, empowerment, orga-
nizational citizenship behavior (OCB), intent to remain within the organiza-
tion, productivity, and self-confidence (Elias, 2008; Podsakoff & Schriesheim,
1985; Randolph & Kemery, 2011). Hard forms of power (e.g., coercive,
reward, and legitimate) are related to greater burnout, absenteeism, and lower
productivity and self-confidence (Elias, 2008; Podsakoff & Schriesheim,
1985; Randolph & Kemery, 2011).
Affect and Power
The emotional aspect of the psychological appraisal process involves employ-
ees’ affective reactions formulated throughout the appraisal process, which
stem from cognitive perceptions of the work events (L. A. James & James,
1989; Lazarus, 1991) and are likely related to employees’ overarching sense of
feelings of well-being across various life domains. In this study, we specifically

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


364 Zigarmi, Roberts, Randolph

investigate work affect—not general well-being—and we define the term affect


as “a phenomenological state of feeling, usually described in terms of emo-
tions such as sad, happy, and enthusiastic, etc.” (Thoresen, Kaplan, Barsky,
Warren, & De Chermont, 2003, p. 914). Such states are lasting affective sup-
positions generated throughout the process of formulating logical conclusions,
and are related to the impact the work events have on one’s general sense of
welfare (cf. L. A. James & James, 1989; Lazarus & Folkman, 1984).
Over the past four decades, mood or feelings have been measured using
standard acceptable instruments (e.g., the Job Affect Scale, Brief, Burke,
George, Robinson, & Webster, 1988; the Positive and Negative Affect Schedule
[PANAS], Watson, Clark & Tellegen, 1988). Two recent meta-analyses (Ng &
Sorensen, 2009; Thoresen et al., 2003) found strong evidence that both posi-
tive and negative affect were related to the nature of the object on which the
respondent was directed to focus. Specifically, both studies maintained that
stronger correlations were probable with positive affect and various positive
valence outcome measures such as job satisfaction, organizational commit-
ment, and personal accomplishment, while negative affect often correlated
with negative valence concepts such as emotional exhaustion, depersonaliza-
tion, and work stress. Applying this logic to the current study, different forms
of power may generate affect of different valence. If respondents view coercive
power and legitimate power as negative valence concepts, for instance, then
these forms of power will likely be inversely related to positive affect.
Citing six studies, Podsakoff and Schriesheim (1985) found that coer-
cive power, reward power, or legitimate power tend to produce either nega-
tive or nonsignificant correlations with satisfaction with supervisor. They also
cited clear and unambiguous positive correlations with expert and referent
power and satisfaction with supervisor. Although no empirical studies could
be found correlating power with specific affect measures, some studies have
suggested relationships between power and constructs that may be affective in
nature such as job satisfaction, organizational commitment, and psychologi-
cal climate (Conger & Kanungo, 1988; Randolph & Kemery, 2011; Thomas
& Velthouse, 1990; Ward, 1998). Also, in a series of experiments, Lammers,
Galinsky, Gordijn, and Otten (2012) found empirical evidence for the connec-
tion between power and heightened social distance. Thus, emerging studies
are laying the groundwork for further investigation of how supervisors’ use of
power may influence subordinates’ affect while on the job.
Furthermore, previous authors’ classifications of hard (coercive, reward,
legitimate) versus soft (referent, expert) power refer to the amount of free-
dom subordinates have in responding to the type of power exerted by their
manager (Pierro, Cicero, & Raven, 2008; Raven et al., 1998). Given that hard
types of power may require compliance in the context of lower-level auton-
omy, it is anticipated that the use of hard power will be more likely to result
in negative feelings in employees than the use of soft power. Therefore, we
propose the following:

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


Employees’ Perceived Use of Leader Power and Implications 365

HYPOTHESIS 1: Coercive, reward, and legitimate power will each negatively correlate
with positive affect and positively correlate with negative affect.

HYPOTHESIS 2: Expert and referent power will each positively correlate with positive
affect and negatively correlate with negative affect.

Work Intentions and Affect


Research has shown, over the past several decades, that intentions reliably
predict behavior (e.g., Ajzen & Fishbein, 1980; Armitage & Connor, 2001;
Bagozzi, 1992). Recent meta-analyses demonstrate that intentions are an
important concept in the attitude-intention-behavior chain (e.g., Armitage
& Connor, 2001; Webb & Sheeran, 2006). Such findings are widespread in
both the health and social psychology literature (e.g., Bagozzi, 1992; Cooke &
Sheeran, 2004; Gollwitzer & Sheeran, 2006; Webb & Sheeran, 2006).
Work intentions result from the second phase of the appraisal process.
As described earlier, the second phase of appraisal results in the individual
selecting among alternative outcomes and courses of action that will eventu-
ally shape their future (Lazarus, 1991). Work intentions flow from the ques-
tion, “What if anything can be done to change or enhance my feelings of
well-being?” Work intentions are formed to solve realized problems or attain
derived needs and wants related to positive or negative feelings of general
well-being (Bagozzi, 1992; Lazarus & Folkman, 1984). Work intentions are
mental images of the behavior an employee plans to manifest (e.g., Ajzen &
Fishbein, 1980; Bagozzi, 1992).
Intentions have been classified into goals and means (Sheeran, Webb,
& Gollwitzer, 2005). Work intentions are defined as “a set of goal represen-
tations or schema and means representations or schema, formed as a result
of the appraisal process, and which are focused upon meeting the needs
and wants stemming from sense of, or lack of, the appraiser’s well-being”
(Zigarmi & Nimon, 2011, p. 450). In this study, the measurement of work
intentions includes various work-related intents found in the organizational
literature over the past 40 years. Specifically, we examine (a) intent to stay
with the organization, (b) intent to use organizational citizenship behav-
iors, (c) intent to perform, (d) intent to use discretionary effort, and (e)
intent to endorse the organization. Both intentions and affect are related
to significant aspects of employees’ work experiences and organizational
performance.
With regard to employee affect, recent meta-analytic studies reveal that
affect correlates with job attitudes, such as organizational commitment, job
satisfaction, employee citizenship, intent to turnover, and dimensions of burn-
out (e.g., Ng & Sorenson, 2009; Thoresen et al., 2003). Positive affect has
been shown to correspond with higher levels of task performance (Kaplan,
Bradley, Luchman, & Haynes, 2009; Tsai, Chen & Liu, 2007).

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


366 Zigarmi, Roberts, Randolph

Additionally, Roberts and Zigarmi (2014) uncovered sizeable correlations


between work intentions and positive job-related affect, such that employees
reporting more favorable affective states were more likely to intend to be orga-
nizational citizens, to perform, to try hard, to remain with their organization,
and to endorse their company. Based on this preliminary work, and in keep-
ing with the appraisal literature, we hypothesize the following:

HYPOTHESIS 3: The five work intentions will positively correlate with positive affect.

HYPOTHESIS 4: The five work intentions will either be negatively correlated or not
significantly correlated with negative affect.

Power, Affect, and Intentions


Furthermore, social cognitive theory (Bandura, 1986) and research on the
appraisal process (Lazarus & Folkman, 1984; Zigarmi, Nimon, Houson, Witt,
& Diehl, 2009, 2011) suggests that employee affect plays a key role in the
relationship between cognition and intentions. Applying this line of think-
ing to the issue of workplace power, it is reasonable to believe that employ-
ees’ cognitive assessment of their manager’s use of power will influence their
intentions (to perform, to stay with and endorse their organization, to engage
in organizational citizenship behaviors, and to use discretionary effort)—but
that this relationship will be partially mediated by employees’ ratings of their
on-the-job affect. Employees’ positive affect should facilitate the relation-
ship between managerial use of power and desirable work intentions, while
employees’ negative affect should be detrimental to the same power-inten-
tions relationship. It is also reasonable to believe that the relationship between
managerial use of power and employee work intentions may not always be
solely dependent on employee affect; rather, power may also directly influence
employee work intentions to some degree. To the best of our knowledge, no
studies testing these relationships exist to date, so our mediation hypothesis
will remain primarily exploratory and is as follows:

HYPOTHESIS 5: Employees’ positive and negative affect will partially mediate the
relationship between managerial use of power and employee work intentions.

Taken together, our conceptual model generally proposes that percep-


tions of various uses of leader power will shape an employee’s affect and feel-
ings of well-being, which will also influence various work intentions.

Method
This section will begin by describing the participants and procedures involved
in this study, and then descriptions of the primary measures used will be
provided.

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


Employees’ Perceived Use of Leader Power and Implications 367

Participants and Procedure


The study’s target population included an e-mail database from a profes-
sional association that was provided to our research team by an international
training and management consulting company. The e-mail database included
managers and nonmanagers of everyday organizations of different sizes and
industries. There is no way of knowing if participants entered into the study
with former knowledge of power, affect, or intentions. The consulting com-
pany providing access to the target population offered no training program on
the topic of organizational power.
Approximately 25,000 individuals from the consulting company’s e-mail
listserv were sent an electronic invitation to participate in this study and
offered access to past research white papers in return for their time.
Of the 773 people who indicated initial interest in participating in the
study, approximately 5% of respondents who opened the survey link dis-
continued their participation after answering fewer than 1% of the survey
questions, so these cases were dropped prior to analysis. Of the remaining
participants, 66% were female, 53% were born in 1961 or earlier, 85% were
Caucasian, 6% were African American, 3% were Hispanic/Latino, 1% were
Asian, and the remaining 5% were of some other racial/ethnic background.
Eighty percent of respondents had either an undergraduate or master’s degree,
14% reported that they had finished some college or less, and the remaining
6% had completed doctorate degrees. With regard to organizational position,
42% were either employees or associates, 37% were supervisors or managers,
and 21% were at the director level or higher. After data cleaning and using
listwise deletion, the final sample for the study’s primary analyses included
651 respondents (2.6% response rate).

Measures
Power Base Survey (PBS). The Power Base Survey was constructed to
measure French and Raven’s five bases of social power, specifically coercive
power, reward power, legitimate power, referent power, and expertise power.
The PBS, developed in a series of studies, is a 25-item instrument measur-
ing the five aspects of managerial power (Randolph & Kemery, 1989, 1990).
Respondents indicated on a 7-point scale (1 = to a very little extent to 7 = to
a great extent) to what degree their manager engaged in certain activities that
reflected the use of the five power bases.
Randolph and Kemery (2011) recently reported updated alpha coeffi-
cients (N=195) for the five power bases. Reward power (e.g., “influences how
much a pay increase others may receive”) was reported to have an alpha coef-
ficient of .74. Coercive power (e.g., “reprimands people for making mistakes”)
revealed a reported alpha coefficient of .79. Legitimate power (e.g., “uses his/
her position or authority to get people to do their tasks”) revealed an alpha
coefficient .82. Expert power (e.g., “knows a great deal about how to do your

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


368 Zigarmi, Roberts, Randolph

and others’ jobs”) showed an alpha coefficient of .85. Referent power (e.g.,
“commands people’s trust and respect”) showed alpha coefficient of .92. Each
scale exceeded the .70 standard advocated by Nunnally (1978).
Positive and Negative Affect. We have chosen a commonly used measure
of affective well-being, the Positive and Negative Affect Scale, PANAS (Wat-
son et al., 1988), to understand the respondents’ feelings of subjective well-
being when meeting with their managers. The scale consists of a shortened
form of the PANAS using 10 items anchored on a 5-point scale, with 1 = very
slightly or not at all to 5 = extremely). Respondents were given the opportu-
nity to evaluate how well certain words (e.g., strong, enthusiastic, interested,
excited, proud, or upset, hostile, guilty, scared, distressed) described how they
felt about their meetings with their boss. The alpha coefficients for these two
scales in this study were positive affect (.90) and negative affect (.84).
Work Intentions. Work intentions were measured by a shortened form
of the Work Intention Inventory (WII). Using two studies, Zigarmi, Nimon,
Houson, Witt, and Diehl (2012) designed and established validation evidence
for their five scales, which assess different forms of work intentions. The five
scales consistently demonstrated acceptable factor structure and reliability
(Zigarmi et al., 2012). Each subscale featured five items and a 6-point Likert-
type response scale, ranging from 1 = no extent to 6 = to the fullest extent.
Example items are as follows: “I intend to continue to work here because I
believe it is the best decision for me” for intent to stay, “I intend to respect this
organization’s assets” for intent to use organizational citizenship behavior, “I
intend to exert the energy it takes to do my job well” for intent to perform, “I
intend to spend my discretionary time finding information that will help this
company” for intent to use discretionary effort, and “I intend to speak out to
protect the reputation of this organization” for intent to endorse, all with pre-
vious alpha coefficients ranging from .83 to .94 (Zigarmi et al., 2012).

Results
The following section is organized as follows: results from preliminary analysis, a
brief overview of how the hypothesized model was tested empirically, a descrip-
tion of the measurement models run, results from tests for common method
variance, results from structural model and mediation model testing, and an
interpretation of the final model with regard to this study’s proposed hypotheses.
Preliminary Analysis
Significant correlations were found among scale scores of many of the vari-
ables, and the nature of these relationships was as anticipated. As shown in
Table 1, expert, referent, and reward power were positively related (rs ranged
from .077 to .729, p < .05), as were legitimate and coercive power (r = .709,
p <  .001). Relationships between referent and expert power (r  =  .729,
p < .001) and between legitimate and coercive power (r = .709, p < .001) were

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


Table 1. Scale Score Means, Standard Deviations, Reliabilities, and Correlations
Mean SD 1 2 3 4 5 6 7 8 9 10 11 12
1) Reward 4.75 1.33 (.887)
Power
2) Coercive 2.78 1.42 .239** (.886)
Power
3) Legitimate 4.14 1.44 .325** .709** (.882)
Power
4) Expert 4.60 1.30 .096* –.266** –.248** (.865)
Power
(5) Referent 4.86 1.56 .077* –.509** –.436** .729** (.963)
Power
(6) Positive 3.08 1.02 .122** –.363* –.338** .619** .686** (.905)
Affect
(7) Negative 1.46 .63 .134** .569** .466** –.376** –.489** –.350** (.842)
Affect
(8) Intent 3.85 1.06 .141** .040 .025 .135** .133** .227** –.003 (.683)
to Use
Discretionary
Effort
(9) Intent to 5.31 .81 .097** –.076* –.061 .169** .210** .347** –.166** .389** (.890)
Perform
Employees’ Perceived Use of Leader Power and Implications

(10) Intent to 4.63 1.24 .152** –.130** –.151** .379** .377** .454** –.218** .332** .459** (.951)
Endorse
(11) Intent to 3.57 1.47 .111** –.239** –.275** .439** .463* .556** –.285** .229** .335** .635** (.917)
Stay
(12) Intent to 5.38 .77 .170** –.032 –.020 .226** .223** .314** –.123** .267** .557** .551** .345** (.852)
Use OCB
Note. Cronbach’s alpha estimates are in parentheses on the diagonal. n = 727, pairwise deletion; reward power figures shown are for the final, 3-item measure.
*Correlation is significant at the .05 level.

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


369

**Correlation is significant at the .01 level.


370 Zigarmi, Roberts, Randolph

high, but these results are in line with theory and previous empirical work;
Randolph and Kemery (2011) reported similarly strong relationships between
the same subscale pairs (r = .69 and r = .57, ps < .01, respectively). Positive
and negative affect were negatively correlated at moderate strength (r = –.350,
p < .001), which supported the orthogonality of the affect measures. All work
intention variables were positively and significantly correlated (rs ranging
from .229 to .635, ps < .001).
Hypothesized Model
This study examined relationships between employees’ cognitive perceptions
of managerial power and affect and the resultant connection between affect
and work intentions. Additionally, the mediating role of both types of affect
on the relationship between power and work intentions was explored through
the examination of direct paths from power to intentions. Confirmatory factor
analysis (CFA) and structural equation modeling (SEM) in EQS were used to
test underlying measurement model fit to the data prior to evaluating subse-
quent structural models.
Measurement Models
CFA was used to evaluate the fit of 12 different measurement models prior to
structural model testing. Overall, the quality of the measurement of power,
affect, and intentions was tested prior to determining overall measurement
model fit for the study’s 12 latent variables. First, a five-factor model was run
with all observed power items loading onto their respective latent variable,
but acceptable fit was not found (χ2[265] = 1536.109, comparative fit index
[CFI] = .91, standardized root mean square residual [SRMR] = .14, root mean
square error of approximation [RMSEA] = .08). Alternative measurement mod-
els for power were run to test for the discriminant validity of the latent fac-
tors for power, particularly to examine potential problems with measurement
model fit that might arise from the two pairs of power latent variables featuring
high correlations (i.e., legitimate and coercive power, and referent and expert
power). A four-factor model specified legitimate and coercive power items to
load onto a single factor with all other observed power items loading onto their
respective factors, and it was compared to the same five-factor model described
earlier. The same was done for referent and expert power. In both cases, the
five-factor model fit the data significantly better than the four-factor models
(Δχ2[4] = 422.466, p < .001 and Δχ2[4] = 1982.835, p < .001, respectively).
A one-factor measurement model was also run, such that all observed
power items loaded onto a single power latent variable. As expected, the one-
factor power model fit significantly worse than the five-factor power model
(Δχ2[10] = 4529.196, p < .001). Thus, the issue with measurement model fit
for power was not due to misspecification of observed items to latent vari-
ables, but rather due to poor loadings (well below .20) of two reward power
observed items onto their designated reward power latent variable. Additional

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


Employees’ Perceived Use of Leader Power and Implications 371

analyses suggested that the reliability of the reward power scale could be nota-
bly improved from .67 to .89 by dropping the two poorly loading items. It
was not our preference to eliminate the two questionable reward power items;
however, we ultimately removed these two items from the analysis because
doing so was necessary to achieve acceptable measurement model fit and
subscale reliability for reward power, and the resultant five-factor model for
power demonstrated acceptable fit to the data (χ2[220] = 753.759, CFI = .96,
SRMR = .06, RMSEA = .06).
Some authors have previously grouped the various forms of power into
broader categories. For example, Raven et al. (1998) grouped different types
of power into hard and soft power bases. Randolph and Kemery (2011) more
broadly referred to expert and referent power as personal power, and legiti-
mate, reward, and coercive power as position power. Lunenberg (2012) com-
bines these five power bases the same way but labels Randolph and Kemery’s
position power as organizational power. In response, we tested for the exis-
tence of second-order individual power and systemic power latent variables in
the measurement model. Although reasonable fit for this hierarchical model
was found (χ2[224] = 837.145, CFI = .96, SRMR = .09, RMSEA = .06), the
five-factor first-order model for power still fit the data significantly better
(Δχ2[4] = 83.386, p < .001) and was thereby accepted as the best measure-
ment model for power.
Two models were run to assess the measurement model fit for affect.
A two-factor model specified all observed affect items loading onto their
respective latent affect variables, either positive or negative. Then a one-
factor model allowed all observed affect items to load onto a single affect
latent variable. As anticipated, the two-factor model for affect fit significantly
better than the one-factor model (Δχ2[1] = 1241.742, p < .001). For work
intentions, a five-factor model with all observed intentions variables load-
ing onto their respective factors demonstrated significantly improved fit over
a model where all observed intentions variables loaded onto a single latent
variable (Δχ2[10] = 2951.931, p < .001). Finally, building on CFA results up
to this point, measurement models were tested for all substantive variables
together. A 12-factor model where all latent variables were allowed to corre-
late demonstrated acceptable fit to the data (χ2[1014] = 2339.394, CFI = .95,
SRMR = .05, RMSEA = .04) as well as superior fit compared to a one-factor
model where all observed items in the study were specified to load onto a
single latent factor (Δχ2[66] = 14654.355, p < .001).
Common Method Variance
Because self-report data collected from a single source may run the risk of
common method bias as indicated by, for instance, the inaccurate inflation of
relationships between substantive constructs (Organ & Ryan, 1995; Spector,
2006), various approaches to evaluating common method variance have been
proposed (Podsakoff, MacKenzie, & Podsakoff, 2012). We used a recently

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


372 Zigarmi, Roberts, Randolph

developed technique by Williams, Hartman, and Cavazotte (2010) to test for


problematic common method variance in our study.
The approach proposed by Williams et al. (2010) compares a series of
five models in CFA to evaluate the degree to which marker variables may be
problematically related to substantive variables in the study. Method variance
bias is not significant among model relationships if Method-R does not dem-
onstrate improved fit over the most relevant fourth model (either Method-C
or Method-U).
We created a latent marker variable using respondents’ survey start date/
time and their Internet Protocol address, neither of which was theoretically
related to the substantive variables in the study. As shown by Table 2, running
the five models in accordance with Williams et al. (2010) indicated that the
Method-U model demonstrated superior fit over the Method-C model (Δχ2
[47] = 270.11, p < .001), suggesting that the model’s method factor loadings
should not be assumed to be equal but rather vary in magnitude. Furthermore,
our Method-R fit the data significantly worse than our Method-U model (Δχ2
[66] = 390.34, p < .001), which provided evidence that the relationships in
our model were not adversely affected by common method variance.
Structural Models and Mediation Testing
For the structural models, respondent age, gender, race/ethnicity, and manage-
rial status were entered as control variables. In all cases, the fit of the data to
the model, significance/direction/magnitude of the structural paths, and final
model interpretation did not change substantially. Therefore, results without
control variables will be offered here for simplicity.
Table 3 summarizes overall fit statistics and mediation testing for compet-
ing structural models. Disturbance terms for all work intentions were specified
to covary, as were the disturbance terms for positive and negative affect. No
error terms for observed items were allowed to covary. Our complete media-
tion model fit the data well (χ2[1039] = 2160.66, CFI = .95, SRMR = .06,
RMSEA = .04), but competing models were run to further evaluate the degree
to which positive and negative affect may be partial mediators of the associa-
tion between power and work intentions (L. R. James, Mulaik, & Brett, 2006).
Competing partial mediation models were analyzed (Models 2–6), such that
each model estimated direct paths from one of the power latent variables to
all five work intentions. Of the five partial mediation models tested, all but
one (Model 6) demonstrated superior fit to the data compared to the complete
mediation model. Because Models 2–5 each indicated three to four significant
direct paths, a final structural model (Model 7) was run that specified all
14 significant direct paths uncovered in Models 2–5.Model 7 fit significantly
better than the complete mediation model (Δχ2[14] = 47.47, p < .001) and
significantly better than partial mediation Models 2–6 (Δχ2[9] = 26.85 to
38.81, all ps < .001). Thus, Model 7 was accepted as final. All standardized
path coefficients for Model 7 are presented in Figure 1.

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


Table 2. Comparison of Measurement Models—Testing for Common Method Variance
Model Description (Number) χ2 df CFI Stand RMR RMSEA
CFA Model (1) 2186.22 1097 0.96 0.05 0.04
Baseline Model (2) 4905.81 1113 0.86 0.77 0.07
Method-C Model (3) 4859.80 1112 0.86 0.77 0.07
Method-U Model (4) 4603.99 1065 0.87 0.77 0.07
Method-R Model (5) 4994.33 1131 0.86 0.78 0.07
ΔModels: Chi-square Comparisons Δχ2 Δdf Sig
Baseline vs. Method-C 31.72 1 p < .001
Method-C vs. Method-U 270.11 47 p < .001
Method-U vs. Method-R 390.34 66 p < .001
Model Description
CFA Model (1) Freely estimates all parameters while specifying correlations among all variables, and it sets method factor
loadings to zero
Baseline Model (2) The method factor loadings are still set to zero, correlations among substantive variables and the marker
Employees’ Perceived Use of Leader Power and Implications

are set to zero, and the factor loadings of observed items onto the marker variable are specified to be their
respective loadings from the first model
Method-C Model (3) Reenters the marker factor’s loadings while constraining them as equal
Method-U Model (4) Allows the marker factor’s loadings to be unconstrained
Method-R Model (5) Exactly the same as the better fitting model between Method-C and Method-U, except latent variable
correlations are constrained to those provided by the second model

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


373
374 Zigarmi, Roberts, Randolph

Table 3. Comparison of Structural Equation Models


Model Δχ2 (Compared
Number Model Description χ2 df CFI SRMR RMSEA to Model 1)
1 Complete 2160.66 1039 0.95 0.06 0.04
Mediation Model
2 Mediation - All 2140.04 1034 0.95 0.05 0.04 20.63, df = 5,
Direct Paths from p < .01
Reward Power
3 Mediation - All 2145.79 1034 0.95 0.05 0.04 14.87, df = 5.
Direct Paths from p < .05
Coercive Power
4 Mediation - All 2148.53 1034 0.95 0.05 0.04 12.13, df = 5.
Direct Paths from p < .05
Legitimate Power
5 Mediation - All 2141.38 1034 0.95 0.06 0.04 19.28, df = 5.
Direct Paths from p < .01
Expert Power
6 Mediation - All 2151.99 1034 0.95 0.06 0.04 8.67, df = 5.
Direct Paths from p < .05
Referent Power
7 Final - 5 2113.19 1025 0.96 0.05 0.04 47.47, df = 14.
Significant Direct p < .001.
Paths
Note. df = degrees of freedom, CFI = Comparative Fit Index, SRMR = Standardized Root Mean
Square Residual, RMSEA = Root Mean Square Error of Approximation

Final Model Interpretation


Of the 10 relationships proposed in Hypotheses 1 and 2, which addressed
the connection between different types of managerial power and positive and
negative affect, 6 were supported by significant path coefficients in the direc-
tions anticipated. Reward power, expert power, and referent power positively
and significantly predicted positive affect (βs = .073, .357, and .367, respec-
tively, ps < .05); employees influenced by these constructive types of power
were more likely to feel favorable emotions at work. Legitimate power was
negatively related to positive affect (β = –.120, p < .05), and expert power
demonstrated an inverse relationship with negative affect (β = –.260, p < .05).
Thus, employees perceiving their managers’ increased use of legitimate power
were more likely to have lower levels of positive affect, and those with manag-
ers exercising greater expert power reported a lower degree of negative affect.
The magnitude of the path coefficient between coercive power and negative
affect was particularly notable, as higher levels of coercive power led to sig-
nificantly increased amounts of negative affect (β = .535, p < .05). It is worth
mentioning that the lack of a significant path between legitimate power and
negative affect should be interpreted with caution, as we would have expected

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


Employees’ Perceived Use of Leader Power and Implications 375

Figure 1. Final Model*

a strong positive relationship in accordance with theory. Given the relation-


ship between legitimate power and negative affect in this study’s scale score
correlation matrix (r = .466, p < .001) and follow-up regression analyses, it
was determined that this path coefficient in the model may be unstable due
to multicollinearity (Grapentine, 2000; Jagpal, 1982) between legitimate and
coercive power.
Seven of the 10 relationships proposed by Hypotheses 3 and 4 were sup-
ported by the final model. Positive affect was moderately and significantly
correlated with all work intentions in the expected direction (βs ranged from
.349 to .487, all ps < .05), indicating that higher levels of favorable affect
while on the job lead to increased behavioral intentions. Results were more
mixed for the paths between negative affect and work intentions, as only two
of five possible paths were significant and those that were significant were
relatively weak in magnitude. Employees experiencing greater amounts of
negative affect were somewhat less likely to intend to perform and to intend
to use OCB at work (β = –.126 and –.131, respectively, ps < .05).
For Hypothesis 5, some support was found for competing structural mod-
els exploring positive and negative affect as partial mediators of the relationship
between power and work intentions; in the final model, there were five signifi-
cant direct paths between power and intentions. Namely, there were significant,
positive direct paths between reward power and intent to endorse (β = .088,
p < .05) and intent to use OCB (β = .112, p < .05). Additionally, significant and

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


376 Zigarmi, Roberts, Randolph

positive direct paths were found between expert power and intent to endorse
(β = .124, p < .05) and intent to stay (β = .133, p < .05). A significant but nega-
tive coefficient was observed for the direct relationship between expert power
and intent to perform (β = –.153, p < .05), which was surprising since both the
correlation between these variables and the indirect effects through positive and
negative affect were all positive. Follow-up regression results showed that the
coefficient’s sign (i.e., representing the relationship between expert power and
intent to perform) was unstable, depending on the other variables in the model;
it would change from significant and positive to nonsignificant when positive
affect was entered into the model, and the sign would flip to significant and neg-
ative when positive affect was entered in the presence of negative affect. In light
of this instability, the general evidence for a positive relationship between expert
power and intent to perform, and our theoretical expectations for a positive rela-
tionship between these variables, we would warn that this relationship may be
a statistical artifact and should not be interpreted without replication in future
research. In the final model, mediation testing supported partial mediation in five
instances, suggesting that certain relationships between power and work inten-
tions may not be fully understood through the effects of on-the-job affect alone.
Effect decomposition analyses, provided in Table 4, examined the total
indirect effects of individual power variables on each work intention variable
through both types of affect. For reward power, indirect effects were signifi-
cant only for intent to use discretionary effort and for intent to stay. The indi-
rect effect of coercive power on intentions through affect was significant only
for intent to use OCB. However, indirect effects of expert and referent power
were significant for all work intentions, supporting the mediating role of posi-
tive and negative affect in these relationships.

Table 4. Total Indirect Effects of Power on Work Intentions Through


Positive and Negative Affect
Power Variables
Reward Coercive Legitimate Expert Referent
Intention Variables Power Power Power Power Power
Intent to Use 0.03 ns ns 0.12 0.14
Discretionary Effort
Intent to Perform ns ns ns 0.21 0.19
Intent to Endorse ns ns ns 0.16 0.15
Intent to Stay 0.03 ns ns 0.18 0.17
Intent to Use OCB ns –0.09 ns 0.16 0.14
Notes. Only standardized coefficients for the significant indirect effects are shown.
ns = not significant at the .05 level. Values shown represent the total indirect effect for
the intention variable on the power variable through both AIM-J positive and AIM-J
negative.

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


Employees’ Perceived Use of Leader Power and Implications 377

Discussion
Positive affect was directly and significantly correlated to all five intentions,
which replicates conclusions drawn from previous work by Roberts and
Zigarmi (2014). Taken together, the notable connection between positive
affect and work intentions suggests that a sense of positive affect is critical for
employees demonstrating productive intentions on the job. In contrast, when
employees experienced lower positive affect and manifested higher levels of
negative affect, their intent to remain, intent to endorse, and intent to use dis-
cretionary effort were not significantly influenced. However, under the same
conditions, employees’ intent to perform and intent to use organizational citi-
zenship behaviors diminished. The significant effects between negative affect
and intention were not as strong as those found between positive affect and
intentions; however, when employees experience negative affect, they may be
somewhat more likely to hold their performance in abeyance and lower their
concern for the welfare of the whole organization.
According to the final structural model, expert power, referent power,
and reward power (though only a small effect was found for the latter) tended
to lead to positive affect, while legitimate power was negatively correlated
with positive affect. Although the path between coercive power and positive
affect was not significant, subscale correlations and additional analyses pro-
vided evidence that the lack of a significant and negative path coefficient was
attributable to a suppressor effect due to the high correlation between coercive
and legitimate power. In practice, this implies that leaders who are interested
in developing a positive feelings of well-being in their direct reports should
use expert power and referent power with their direct reports, while deem-
phasizing the use of legitimate and coercive power as much as possible.
Figure 1 indicates that coercive power strongly correlates with nega-
tive affect and increases the possibility of reduced intentions to perform and
intentions to use organizational citizenship behaviors from their employees.
A second consideration is that when leaders increase their use of legitimate
power, they reduce their employees’ positive feelings and lessen the possi-
bility of all five of the intentions occurring from their followers. These two
findings are commensurate with various studies (Conger & Kanungo, 1988;
Randolph & Kemery, 2011; Thomas & Velthouse, 1990). As pointed out by
Randolph and Kemery, “A manager who is trying to delegate authority would
not want to rely on coercion to get people to take on additional responsibility,
nor would the manager want to exert legitimate power when trying to reduce
employee reliance on the manager” (2011, p. 98). We can also conclude that
the use of referent and expert power somewhat increases the probability of
greater levels of employee positive affect and a possible increase in employee
discretionary effort, higher levels of performance, endorsement of the organi-
zation, more frequent organizational citizenship behaviors, and lower inten-
tions to turnover. This is somewhat in keeping with the meta-analysis study

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


378 Zigarmi, Roberts, Randolph

done by Podsakoff and Schriesheim (1985), who found that withdrawal inten-
tions were negatively correlated with referent and expert power. They also
found that reward, coercive, referent, and legitimate power were unrelated
to employees’ perceived support of their supervisor and their work commit-
ment. On the other hand, expert power was moderately correlated with sub-
ordinates’ “satisfaction with supervisor” (Podsakoff & Schriesheim, 1985).
Mossholder et al. (1998) showed moderate correlations between expert and
referent power and job satisfaction (rs = .35 and .35, respectively).
The current study suggests that managers can clearly impact employee
affect by their use of power. Bases of power like coercive and legitimate can
result in negative affect and ultimately in a lack of positive intentions toward
the organization, while expert and referent power result in positive affect and
thus in enhanced positive intentions. More research is needed to expand our
understanding of these key managerial effects on employees.
Study Limitations and Future Directions
This study is not without limitations. Employee perceptions served as the only
source of information for this research. Although self-ratings are arguably an
ideal source for data on subjective psychological processes pertaining to affect
and intentions, that may not always be the case for ratings of managerial use
of power. While employee perceptions of power use was of primary interest
in this study and measured accordingly, it is possible that managers’ ratings
of their own power could provide useful additional information about how
power is enacted and how it may affect others in relational exchanges at work.
Thus, collecting data from managers may be helpful in later work.
Also, the cross-sectional nature of the data did not allow investiga-
tion into the causal ordering of the psychological processes studied. Future
research could gather longitudinal or qualitative data to better understand how
the psychological processes in question function in real time. Additionally,
because respondents were invited to participate from an e-mail list housed by
a single organization, results may not be generalizable to other populations.
The sample analyzed was predominantly Caucasian, and the findings may
not apply to individuals of other racial/ethnic backgrounds. Cross-cultural or
cross-industry research examining the connection between power, affect, and
work intentions may enable more to be learned about how the constructs may
or may not manifest differently in other settings.
Other limitations to this work include the measurement of intent to
use discretionary effort, which involved a three-item scale that fell at .68,
or just below the widely accepted .70 minimum alpha reliability coefficient
(Cronbach, 1951). Additionally, as part of the measurement model testing nec-
essary prior to running SEM, we tested the existence of a higher-order latent
variable for power; from this, two points are worth mentioning. Although
models that specified power as two higher-order variables were not the best
fitting measurement models in this study overall, the generally sufficient fit of

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


Employees’ Perceived Use of Leader Power and Implications 379

these models raises important measurement questions regarding the construct


of power for future study in the workplace. Authors have loosely classified dif-
ferent types of power (Lunenberg, 2012; Randolph & Kemery, 2011; Raven
et al., 1998), and initial work has shown substantial correlations between
certain kinds of power (Randolph & Kemery, 2011). All of this raises the
question as to whether or not measuring some types of power may best be
accounted for by modeling them at a higher level of abstraction. Specifically,
are there latent aspects to certain kinds of power that can be more optimally
understood under a new classification system for power types? As recent work
has extended the study of power beyond French and Raven’s (1959) original
5 power bases to include 11 different types of power (Raven, 1993; Raven
et al., 1998) the debate over definitions and forms of power continues. Thus,
exploring and improving the measurement of power as a construct may be a
promising area for future work.
Implications for HRD Research
Going forward, it is recommended that other types of managerial power be
researched for their likely connection to employee affect and work intentions.
Or, with regard to work intentions, future research could include intentions
that are malicious or counterproductive—that is, not only constructive in
nature, as the latter was the focus of this study. It would also be interesting
to investigate the role of negative affect in the relationship between coercive
or legitimate power and harmful or deviant work intentions. Expanding a
future focus to the organizational level, another recommendation for research
could be to examine the impact of systemic policies and procedures, which
are not exclusive to individual management behavior but are also linked to
employees’ psychological appraisal processes inclusive of affect and intentions.
Furthermore, expanding research from the realm of doctoral student supervi-
sion by Doloriert, Sambrook, and Stewart (2012) into the context of HRD, it
is recommended that future work be conducted exploring the degree to which
employees actively manage their emotions in response to supervisorial power.
It is likely the case that the research implications underlying this work are more
pressing than corresponding practice-in-use implications. However, we still
believe it is important to next briefly address practical implications of this work.
Implications for HRD Practice
One practical takeaway from this article is that HRD practitioners should real-
ize that various forms of power shape the structure and quality of role rela-
tionships at work, and this knowledge should inform the design of future
leadership and team training. The results of this study as well as others (e.g.,
Podsakoff & Schriesheim, 1985; Simmons & Roberson, 2003; Van Dijke,
De Cremer, & Mayer, 2010) demonstrate that authority and use of various
forms of “hard” power drastically shape employee perceptions of organiza-
tional fairness, which in turn influences organizational outcomes such as

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


380 Zigarmi, Roberts, Randolph

employee turnover rates, satisfaction with boss, and levels of performance.


Various forms of participative management or empowerment calls for an equal
distribution of power in service of those who will be empowered. Leaders
must be educated to identify the different forms of power that will encourage
autonomy for those who must implement strategic decisions.
Another strategy for HRD practitioners is to consider the examination
of policies and procedures, and the managers’ use of various forms of power
these policies and procedures imply—particularly with regard to low-stakes
versus high-stakes work contexts. For example, in an industry that mandates
high levels of employee work safety, such as the airlines, coalmines, or nuclear
industry, the use of coercive power might be more legitimized in the minds
of organizational members because the critical concern for safety should out-
weigh unfavorable repercussions of employees experiencing negative affect.
As the results of this study suggest, there are benefits to exercising certain
forms of power, such as greater employee engagement, greater organizational
citizenship behavior, and greater use of employee discretionary effort. This
strong implication of this study involves HRD making the business case for
awareness of when coercive, legitimate, and reward power should be judi-
ciously used for specific purposes such as boundary formation and manage-
ment. Alternatively, expertise, referent, and informational power may be better
used to increase employee attributions of meaningful work, build pride in
ownership of daily tasks, and inspire creativity through autonomy. Given the
importance of these aspects of work to knowledge workers and Generation Y
employees in particular, a closer examination of managerial use of power is
warranted in the future.
Another notable implication for HRD practitioners concerns the poten-
tial value of developing curriculum addressing the use of political skills for
all levels of leadership in their organizations. Some research has shown that
valuable political skills including social astuteness, interpersonal influence,
network ability, and apparent sincerity can be developed and can make a help-
ful difference in the political environment of an organization (Ferris et al.,
2007). Recently, the Political Skills Inventory (PSI) (Ferris et al., 2005) has
been developed and validated, and it has been tied to organizational outcomes
such as job performance and effectiveness ratings. Perhaps HRD practitioners
could use similar concepts or tools to improve curriculum offerings made
available to their corporate executives.

References
Aguinis, H., Nesler, M. S., Quigley, B. M., & Tedeschi, J. T. (1994). Perceptions of power: A
cognitive perspective. Social Behavior and Personality, 22, 377–384.
Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting social behavior. Englewood
Cliffs, NJ: Prentice Hall.
Armitage, C. J., & Connor, M. (2001). Efficacy of the theory of planned behavior: A meta-
analytic review. British Journal Social Psychology, 40, 471–500.

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


Employees’ Perceived Use of Leader Power and Implications 381

Bagozzi, R. P. (1992). The self-regulation of attitudes, intentions, and behavior. Social Psychology
Quarterly, 55, 178–204.
Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Upper Saddle
River, NJ: Prentice Hall.
Brief, A. P., Burke, M. J., George, J. M., Robinson, B. S. & Webster, J. (1988). Should nega-
tive affectivity remain an unmeasured variable in the study of job stress? Journal of Applied
Psychology, 73, 193–198. doi:0021-9010/88/$00.75
Byrd, M. (2008). Negotiating new meanings of leader and envisioning culturally informed
theories of developing African-American women in leadership roles: an interview with
Patricia S Parker. Human Resource Development International, 11(1), 101–107. doi: 10.
1080/13678860701782477
Callahan, J. L., Whitener, J. K. & Sandlin, J. A. (2007). The art of creating leaders: Popular
culture artifacts as pathways for development. Advances in Developing Human Resources, 9,
146–165. doi:10.1177/15234223036298856
Clegg, R., Courpasson, D., & Phillips, N. (2011). Power and organizations. Thousand Oaks, CA:
Sage.
Conger, J. A., & Kanungo, R. N. (1988). The empowerment process: Integrating theory and
practice. Academy of Management Review, 13, 471–482. doi:10.5465/AMR.1988.430-6983
Cooke, R., & Sheeran, P. (2004). Moderation of cognition-intention and cognition-behavior
relations: A meta-analysis of properties of the variables from the theory of planned behavior.
British Journal of Social Psychology, 43, 159–186.
Cronbach, L. J. (1951). Coefficient alpha and the internal structure of tests. Psychometrika, 16,
297–334.
DeCelles, K. A., DeRue, D. S., Margolis, J. D., & Ceranic, T. L. (2012). Does power corrupt or
enable? When and why power facilitates self-interested behavior. Journal of Applied Psychology,
97(3), 681–689. doi:10.1037/a0026811
Doloriert, C., Sambrook, S., & Stewart, J. (2012). Power and emotion in doctoral supervi-
sion: Implications for HRD. European Journal of Training and Development, 36(7), 732–750.
doi:10.1108/03090591211255566
DuBrin, A. J. (2009). Political behavior in organizations. Thousand Oaks, CA: Sage.
Elias, S. (2008). Fifty years of influence in the workplace: The evolution of the
French and Raven power taxonomy. Journal of Management History, 14, 267–283.
doi:10.1108/17511340810880634
Farmer, S. M., & Aguinis, H. (2005). Accounting for subordinate perceptions of supervi-
sory power: An identity-dependence model. Journal of Applied Psychology, 90, 1069–1083.
doi:101037/0021-9010.90.6.1069
Ferris, G. R., Treadway, D. C., Perrewe, P. A. L., Brouer, R. L., Douglas, C. & Lux, S.
(2007). Political skill in organizations. Journal of Management, 33, 290–320. doi:10.
1177/0149206307300813
Ferris, G. R., Treadway, D. C., Kolodinsky, R. W., Hochwarter, W. A., Kacmar, C. J., Douglas, C.,
& Frink, D. D. (2005). Development and validation of the political skill inventory. Journal of
Management, 31, 126–152. doi:10. 1177/1049206304271386
French, J., & Raven, B. H. (1959). The bases of social power. In D. Cartwright (Ed.), Studies in
social power (pp. 150–167). Ann Arbor, MI: Institute for Social Research.
Germain, M., & Tejeda, M. J. (2012). A preliminary exploration of the measurement of expertise:
An initial development of a psychometric scale. Human Resource Development Quarterly, 23,
203–232. doi:10.1002/hrdq. 21134
Gollwitzer, P. M., & Sheeran, P. (2006). Implementation intentions and goal achievement: A
meta-analysis of affect and processes. Advances in Experimental Social Psychology, 38, 69–119.
doi:10.1016/S0065-2601(06)38002-1
Grapentine, T. (2000). Path analysis vs. structural equation modeling. Marketing Research, 12,
12–20.

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


382 Zigarmi, Roberts, Randolph

Gubbins, M. C., & Garavan, T. N. (2005). Studying HRD practitioners: A social capital model.
Human Resource Development Review, 4, 189–218. doi:10. 1177/1534484305275769
Haugaard, M., & Clegg, S. (2012). Power and organizations. Thousand Oaks, CA: Sage.
Hopfl, H. M. (1999). Power, authority, and legitimacy. Human Resource Development International,
2(3), 217–234. doi: 10. 1080/13678869900000024
Jagpal, H. S. (1982). Multicollinearity in structural equation models with unobservable variables.
Journal of Marketing Research, 19, 431–439.
James, L. A., & James, L. R. (1989). Integrating work environment perceptions: Explorations into the
measurement of learning. Journal of Applied Psychology, 74, 739–751. doi:0021-9010/89/$00.75
James, L. R., Mulaik, S. A., & Brett, J. M. (2006). A tale of two methods. Organizational Research
Methods, 9, 233–244.
Kaplan, S., Bradley, J. C., Luchman, J. N., & Haynes, D. (2009). On the role of positive and neg-
ative affectivity in job performance: A meta-analytic investigation. Journal of Applied Psychology,
94, 162-176. doi:10.1037/a001-3115
Keller, D. (2007). Leading on top of the world: Lessons from into thin air. Advances in Developing
Human Resources, 9, 166–182. doi:10.1177/1523422306298857
Kochery, T. S. (1993). Conflict versus consensus: Processes and their effect on team decision-
making. Human Resource Development Quarterly, 4, 185–191. doi:10.1002/hrdq.392-004-0208
Krantz, J. (2006). Leadership, betrayal and adaptation. Human Relations, 59, 221–249.
doi:10.1177/0018726706062733
Lammers, J., Galinsky, A. D., Gordijn, E. H., & Otten, S. (2012). Power increases social distance.
Social Psychological and Personality Science, 3(3), 282–290.
Lammers, J., & Stapel, D. A. (2009). How power influences moral thinking. Journal of Personality
and Social Psychology, 97(2), 279–289. doi:10.1037/a0015437
Lazarus, R. S. (1991). Emotion and adaptation. New York, NY: Oxford University Press.
Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and coping. New York, NY: Springer.
Lord, R. G., & Emirch, C. G. (2001). Thinking outside the box by looking inside the box:
Extending the cognitive revolution and leadership research. Leadership Quarterly, 11, 551–
579. doi:10.1016/S1048-9843(00)00060-6
Lord, R. G., & Kernan, M. C. (1987). Scripts as determinants of goal commitment. Academy of
Management Review, 12, 265–277. doi:10.5465/AMR.1987.430-7831
Lunenberg, F. C. (2012). Power and leadership: An influence process. International Journal of
Management, Business, and Administration, 15, 1–9.
Maner, J. K., & Mead, N. L. (2010). The essential tension between leadership and power: When
leaders sacrifice group goals for the sake of self-interest. Journal of Personality, 99(3), 482–497.
doi:10.1037/a0018559
Montesino, M. U. (2002). A descriptive study of some organizational-behavior dimensions at
work in the Dominican Republic: implications for management development and training.
Human Resource Development International, 5(4), 393–409. doi: 10.1080/13678860110059366
Mossholder, K. W., Bennett, N., Kemery, E. R., & Wesolowski, M. A. (1998). Relationships
between the bases of power and work reactions: The mediational role of procedural justice.
Journal of Management, 24(4), 533–552. doi:10.1177/014920639802400404
Ng, T. W. H., & Sorensen, K. L. (2009). Dispositional affectivity and work-related out-
comes: A meta-analysis. Journal of Applied Social Psychology, 39(6), 1255–1287.
doi:10.1111/j.1559-1816.2009.00481.x
Nunnally, J. C. (1978). Psychometric theory (2nd ed.). New York, NY: McGraw-Hill.
Organ, D. W., & Ryan, K. (1995). A meta-analytic review of attitudinal dispositional predictors
of organizational citizenship behavior. Personnel Psychology, 48, 775–802.
Pfeffer, J. (1992). Managing with power: Politics and influence in organizations. Boston, MA: Harvard
Business School Press.
Pfeffer, J. (2011). Power: Why some people have it—and others don’t. New York, NY: HarperCollins.
Pierro, A., Cicero, L., & Raven, B. H. (2008). Motivated compliance with bases of social power.
Journal of Applied Psychology, 38(7), 1921–1944.

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


Employees’ Perceived Use of Leader Power and Implications 383

Pitesa, M., & Thau, S. (2013). Masters of the universe: How power and accountability influ-
ence self-serving decisions under moral hazard. Journal of Applied Psychology, 98(3), 550–558.
doi:10.1037/a0031697
Podsakoff, P. M., MacKenzie, S. B., & Podsakoff, N. P. (2012). Sources of method bias in social
science research and recommendations on how to control it. Annual Review of Psychology, 65,
539–569. doi:10.1146/annurev-psych-120710-100452
Podsakoff, P. M., & Schriesheim, C. A. (1985). Field studies of French and Raven’s bases of
power: Critique, reanalysis, and suggestions for future research. Psychological Bulletin, 97(3),
387–411. doi:10.1037/0033-2909.97.3.387
Randolph, W. A., & Kemery, E. R. (1989). Toward the construct validity of a measure of French
and Raven’s (1959) power base typology. Eastern Academy of Management Proceedings, 1–3.
Randolph, W. A., & Kemery, E. R. (1990). An empirical comparison of two recent scales
for operationalizing French and Raven’s bases of power. Eastern Academy of Management
Proceedings, 120–123.
Randolph, W. A., & Kemery, E. R. (2011). Managerial use of power bases in a model of manage-
rial empowerment practices and employee psychological empowerment. Journal of Leadership
& Organizational Studies 18(1), 95–106. doi:10.1177/1548051810379798
Raven, B. H. (1993). The basis of power: Origins and recent developments. Journal of Social
Behavior and Personality, 7, 217–244.
Raven, B. H., Schwarzwald, J. & Kozlowsky, M. (1998). Conceptualizing and measuring a
power/interaction model of interpersonal influence. Journal of Applied Social Psychology, 28,
307–332. doi:10.1111/j1559-1816.1998.tb01707.x
Roberts, T. P., & Zigarmi, D. (2014). The impact of dispositional cynicism on job-specific
affect and work intentions. International Journal of Psychology, 49(5), 381–389. doi:10.1002/
ijop.12051
Seibert, S. E., Silver, S. R., & Randolph, W. A. (2004). Taking empowerment to the next level: A
multiple-level model of empowerment, performance, and satisfaction. Academy of Management
Journal, 47(3), 332–349. doi:10.1111/j.1744-6570.2012.0125.X
Sheeran, P., Webb, T. L., & Gollwitzer, P. M. (2005). The interplay between goal inten-
tions and implementation intentions. Personality and Social Psychology Bulletin, 31, 87–98.
doi:10.1177/0146167204271308
Simmons, T., & Roberson, Q. (2003). Why managers should care about fairness: The effects of
aggregate justice perceptions on organizational outcomes. Journal of Applied Psychology, 88(3),
432–443. doi:10.1037/0021-9010.88.3.432
Spector, P. (2006). Method variance in organizational research: Truth or urban legend?
Organizational Research Methods, 9(2), 221–232. doi:10.1177/1094428105284955
Thomas, K. W., & Velthouse, B. A. (1990). Cognitive elements of empowerment: An “inter-
pretive” model of intrinsic task motivation. Academy of Management Review, 15, 666–681.
doi:10.546/AMR.1990.431-0926
Thoresen, C. J., Kaplan, S. A., Barsky, A. P., Warren, C. R., & De Chermont, K. (2003). The affec-
tive underpinnings of job perceptions and attitudes: A meta-analytic review and integration.
Psychological Bulletin, 129(6), 914–945. doi:10.1037/0033-2909.129.6.914
Tjosvold, D. (2006). Effects on power concepts and employee performance on manag-
ers’ empowering. Leadership and Organizational Development Journal, 27, 217–234.
doi:10.1108/01437730610657730
Tsai, C., Chen, C., & Liu, H. (2007). Test of a model linking employee positive moods and task
performance. Journal of Applied Psychology, 92, 1570–1583. doi:10.1037/0021-9010.92.6.1570
Van Dijke, M., DeCremer, D., & Mayer, D. M. (2010). The role of authority power in explaining
procedural fairness effects. Journal of Applied Psychology, 95, 488–502. doi:10.1037/a 001-8921
Vince, R. (2014). What do HRD scholars and Practitioners need to know about power, emotion,
and HRD?. Human Resource Development Quarterly, 25(4), 409–420. doi: 10.1002/hrdq.21191
Ward, E. A. (1998). Managerial power bases and subordinates’ manifest needs as influences on
psychological climate. Journal of Business and Psychology, 12(3), 361–378.

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


384 Zigarmi, Roberts, Randolph

Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures
of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology,
54, 1063–1070. doi:0022-3514/88/$00.75
Webb, T. L., & Sheeran, P. (2006). Does changing behavioral intentions engender behavioral
change? A meta-analysis of the experimental evidence. Psychological Bulletin, 132, 249–268.
doi:101037/0033-2909.132.2.249
Williams, L. J., Hartman, N., & Cavazotte, F. (2010). Method variance and marker variables: A
review and comprehensive CFA marker technique. Organizational Research Methods, 13(3),
477–514. doi:10.1177/1094428110366036
Wofford, J. C., & Goodwin, V. L. (1990). Effects of feedback on a cognitive processing and choice
of decision-making style. Journal of Applied Psychology, 75, 603–612. doi:0021-0910/90/$00.75
Yorks, L. (2004). Toward a political economy model of comparative analysis of the role of stra-
tegic human resource development leadership. Human Resource Development Review, 3(3),
189–208. doi:10.1177/1534484304266260
Zigarmi, D., & Nimon, K. (2011). A cognitive approach to work intention: The stuff that
employee work passion is made of? Advances in Developing Human Resources, 13, 447–461.
doi:10.1177/1523422311431152
Zigarmi, D., Nimon, K., Houson, D., Witt, D., & Diehl, J. (2009). Beyond engagement:
Toward a framework and operational definition for employee work passion. Human Resource
Development Review, 8, 300–326. doi:10.1177/153448309338171
Zigarmi, D., Nimon, K., Houson, D., Witt, D., & Diehl, J. (2011). A preliminary field test of
an employee work passion model. Human Resource Development Quarterly, 22(2), 195–221.
doi:10.1002/hrdq.20096
Zigarmi, D., Nimon, K., Houson, D., Witt, D., & Diehl, J. (2012). The work intention inven-
tory: Initial evidence of construct validity. Journal of Business Administration Research, 1, 24–42.
doi:10.5430/jbar.v1n1p24

Dr. Drea Zigarmi, the Director of Research for the Ken Blanchard Companies, has
published five books on leadership, and authored several articles in various journals. Drea
also teaches at the University of San Diego in the Master of Science in Executive Leadership
Program. He holds a bachelor’s degree in biology from Norwich University, a master’s
degree in humanistic education, and an EdD in organizational studies from the University
of Massachusetts.

Dr. Taylor Peyton Roberts conducts research for the Ken Blanchard Companies and is
Managing Partner and Co-Founder of Valencore Consulting, a leadership development
and research firm in San Diego. She holds a B.S. in Psychology from California State
University Long Beach, an M.S. in Industrial/Organizational Psychology from San Diego
State, and a Ph.D. in Leadership Studies from the University of San Diego. She has held
teaching positions for San Diego State University’s Department of Management and for the
University of San Diego’s Department of Leadership Studies.

Dr. W. Alan Randolph is CSX Distinguished Professor at the University of Baltimore. Alan’s
research interests focus on empowerment and global leadership skills. He has published
widely in journals and authored eight books. Alan holds a bachelor’s degree in industrial
engineering from Georgia Institute of Technology, and an MS in business administration
from the University of Massachusetts.

Corresponding author:
Taylor Peyton Roberts can be contacted at tpeyton@sandiego.edu

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq

You might also like