Anda di halaman 1dari 33

QUANTITATIV E STUDY

Exploring Different
Operationalizations of Employee
Engagement and Their
Relationships With Workplace
Stress and Burnout

Paula E. Anthony-McMann, Andrea D. Ellinger, Marina Astakhova,


Jonathon R. B. Halbesleben

Many empirical studies of employee engagement show positive


relationships with desirable work-related outcomes, yet a consistent
understanding of the construct remains elusive (Saks & Gruman, 2014).
We propose that this lack of clarity is leading to an increased risk that
employee engagement is becoming overly generalized and that, as a
consequence, its utility in both theory and practice is compromised. Indeed,
our study of 472 information technology (IT) professionals working
in community hospitals reveals that, even though the two measures of
employee engagement examined in this study are conceptually based
on Kahn’s (1990) needs-satisfaction framework, they are nomologically
different, evidence different predictive properties (with regard to workplace
stress and burnout), and suggest different workplace interventions. As
hypothesized, both measures of employee engagement reveal negative
relationships with workplace stress, and burnout has a mediating effect
on those relationships. Further, the relationships are significantly different,
but these differences are understood only when examining the dimensional
levels of each engagement measure. Our findings also clarify the highly
debated relationship between employee engagement and burnout and
challenge those engagement measures that are conceptually grounded
in a burnout-antithesis framework. Implications and avenues for future
research are presented.

Key Words: burnout, employee engagement, Kahn, measurement,


nomological framework, workplace stress
HUMAN RESOURCE DEVELOPMENT QUARTERLY, vol. 28, no. 2, Summer 2017 © 2016 Wiley Periodicals, Inc.
Published online in Wiley Online Library (wileyonlinelibrary.com) • DOI: 10.1002/hrdq.21276 163
164 Anthony-McMann, Ellinger, Astakhova, Halbesleben

Engagement research has proliferated in the past decade and suggests that
employee engagement is related to many positive work-related outcomes such
as job satisfaction (Shuck, Reio, & Rocco, 2011), job performance (Rich, Lep-
ine, & Crawford, 2010), profitability (Harter, Schmidt, Agrawal, & Plowman,
2013), customer satisfaction, and employee retention (Halbesleben, 2010).
However, it remains difficult to draw precise conclusions from these studies
due to the nomologically different ways in which employee engagement is
measured (Saks & Gruman, 2014). There is little doubt that employee engage-
ment remains a compelling topic for many scholars, but if it is to become a
relevant barometer against which certain organizational decisions are evalu-
ated, then clarity about what is actually being measured is critical.
Employee engagement emerged in the management literature from
Kahn’s (1990) seminal study on personal engagement and disengagement,
but a review of the literature suggests that engagement is now conceptual-
ized by at least four major frameworks and operationalized by at least 10
measurement instruments (Keenoy, 2014; Shuck, 2011). In some cases, dif-
ferent measurement scales reflect subtle differences in the conceptualizations
of engagement, and in other cases, significant differences in their underlying
nomological frameworks—all of which suggest that different measurements
of engagement may actually measure different aspects of engagement. Indeed,
“although it finds its origin in the positive psychology of Kahn (1990), …
[engagement] has, in effect, taken on a life of its own (or, more precisely, a
series of parallel lives)” (Keenoy, 2014, pp. 197–198).
We contend that these ‘parallel lives’ are leading to an increased risk that
the meaning of engagement is becoming elusive, which compromises its util-
ity both in theory and as an actionable phenomenon. This confounds our
understanding of the construct, particularly in relation to other variables, and
makes its operationalization all the more difficult (Shuck, Ghosh, Zigarmi,
& Nimon, 2012). It also promotes continued debate as to the uniqueness of
engagement in comparison to other, more clearly defined, work-related phe-
nomena such as job satisfaction and organizational commitment (Shuck et al.,
2012). Finally, it results in measurements of engagement that are no longer
consistent with Kahn’s (1990) holistic perspective.
Our study suggests that the relationships between employee engage-
ment and certain workplace outcomes depend, in fact, on how the con-
struct of employee engagement is actually measured. We also suggest that
even when different measures of engagement are similarly conceptualized,
we will see different relationships with both predictor and outcome vari-
ables. To explore this, we focused our study on two predictors of employee
engagement that are among the most complex in the management literature:
workplace stress and burnout. If engagement’s relationship to key predictors
differs depending on the measurement instrument, then it is reasonable to
assume that these different instruments will have different predictive proper-
ties as well. This means that the utility of the instruments used to measure

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


Exploring Different Operationalizations of Employee Engagement 165

employee engagement is only as good as our understanding of what is actu-


ally being measured.
This study makes significant contributions to the engagement litera-
ture in that it offers a nuanced explanation of some of the differences in how
employee engagement is measured. We evaluate the relationships between
workplace stress, burnout, and engagement through the lens of Kahn’s (1990)
needs-satisfaction–based framework in response to a resurgence in scholarly
interest in this engagement conceptualization (Fletcher & Robinson, 2013).
We also explore some of the tensions in the literature arising from the agree-
ment (or disagreement) of various scholars as to the nature of the relation-
ship between burnout and engagement. Finally, practical implications related
to engagement measurement are provided, which may improve the efficacy
of workforce interventions intended to support the sustained engagement of
employees—one of today’s key human resource–related imperatives.

Theory and Hypothesis Development


This section reviews the literature and theory in support of the development
of the hypotheses and conceptual model tested in this study.
Employee Engagement and Its Measurement
The concept of engagement emerged from the positive psychology move-
ment in which researchers began to focus on understanding the factors that
can lead to and sustain positive human behaviors and the related positive
consequences of those behaviors. Kahn’s (1990) grounded theory study con-
ceptualized personal engagement as “the simultaneous employment and
expression of a person’s ‘preferred self’ in task behaviors that promote con-
nections to work and to others, personal presence (physical, cognitive, and
emotional), and active full role performances” (p. 700). More specifically,
Kahn (1990) posited that individuals are motivated to engage (or disengage)
in response to how they see themselves in specific roles and with respect to
three psychological conditions: meaningfulness, safety, and availability. His
later work emphasized the primacy of psychological safety, predicated on the
presence of positive and trusting interpersonal relationships at work, as key
to understanding how engagement is sustained (Kahn, 2007). Despite the
broad acceptance of Kahn’s conceptualization in the scholarly community and
the thousands of citations his groundbreaking research has generated, four
distinct frameworks of engagement have since emerged: the needs-satisfac-
tion, burnout-antithesis, job satisfaction, and multidimensional frameworks
(Fletcher & Robinson, 2013; Shuck, 2011). Indeed, these frameworks share
some commonalities, most notably that engagement is a motivational state
that depends on an employee’s perception of certain valued resources (Saks &
Gruman, 2014). However a brief review of each framework reveals important
differences among them.

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


166 Anthony-McMann, Ellinger, Astakhova, Halbesleben

Needs-Satisfaction Framework According to Kahn (1990) the presence


of (or absence of) resources, described in terms of psychological conditions,
influence employees to engage (or disengage). The first, meaningfulness, “can
be seen as a feeling that one is receiving a return on investments of one’s full
self in a currency of physical, cognitive, or emotional energy” (pp. 703–704).
A sense of meaningfulness relates to employees feeling valued and worthwhile
and is correlated to (a) challenging, creative, and autonomous task character-
istics; (b) role characteristics of shared expectations, status, and/or influence;
and (c) work interactions with coworkers and clients that are positive and
reflect rewarding interactions (Kahn, 1990, 1992). His second psychological
requirement for engagement, psychological safety, reflects the extent to which
an individual can express his or her preferred self without fear of negative
consequences. Factors that support psychological safety include positive, con-
sistent, and nonthreatening interactions with coworkers and supervisors and
the establishment of shared expectations. The third, psychological availability,
is “the sense of having the physical, emotional, or psychological resources
to personally engage at a particular moment” (Kahn, 1990, p. 714) and is
impacted by physical and emotional energies, insecurities, relationships, dis-
tractions, and outside life interactions.
Numerous researchers have attempted to operationalize Kahn’s (1990)
needs-satisfaction conceptualization. In 2004, May, Gilson, and Harter (2004)
conducted a study that validated all three of Kahn’s psychological precon-
ditions; however, their measurement instrument revealed some reliabil-
ity challenges at the subscale or dimensional level in later testing (Viljevac,
Cooper-Thomas, & Saks, 2012). Rich et al. (2010) drew on Kahn’s theory to
“describe how engagement represents the simultaneous investment of cog-
nitive, affective, and physical energies into role performance” (p. 617). This
work refined earlier operationalizations and found that engagement mediated
the relationship between certain antecedents such as perceived organizational
support and the employee behaviors of task performance and organizational
citizenship behavior (Rich et al., 2010). It also resulted in one of the purest
operationalizations of Kahn’s (1990) original model (Shuck et al., 2012). More
recently, Soane et al. (2012) expanded the operationalization of the needs-sat-
isfaction framework. They acknowledged the importance of intellectual activity
and the role of affect, but added a social dimension, which reflected their belief
that Kahn’s conceptualization of engagement was predicated on the presence
of positive relationships at work. Describing positive, trusting relationships as
resilient, such relationships are possible only in environments in which employ-
ees feel safe to take risks and/or accept personal vulnerabilities (Kahn, 2007).
Further, it develops only when “members join together in meaningful ways to
share information, solve problems, make sense of their experiences and pro-
vide support” (p. 189). The results of the Soane et al. (2012) study revealed
significant relationships between engagement and the outcomes of task perfor-
mance, organizational citizenship behavior, and intent to stay.

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


Exploring Different Operationalizations of Employee Engagement 167

Burnout-Antithesis Framework. Some of the most prolific researchers


in the field of burnout position engagement as the antithesis (or antipode) of
burnout. Their frameworks suggest that, since burnout is the same thing as
disengagement, an engaged individual must not be burned out (Demerouti
& Bakker, 2008; Maslach, Schaufeli, & Leiter, 2001; Shirom, 2005). One of
the most popular of these operationalizes a definition of engagement as “the
positive, fulfilling work-related state of mind that is characterized by vigor,
dedication and absorption” (Schaufeli, Bakker, & Salanova, 2006, p. 702).
However, even the scholars who position engagement as a distinct construct
from burnout, still measure engagement with instruments (e.g., the Utrecht
Work Engagement Scale [UWES]) in which many questions are almost the
exact opposite of questions in the popular burnout measure, the Maslach
Burnout Inventory (MBI) (Schaufeli, Salanova, González-Romá, & Bakker,
2002). In fact, a meta-analytic review of empirical findings from the UWES
and MBI concluded that the UWES may measure the antipode of burnout,
but not necessarily the construct of engagement (Cole, Walter, Bedeian, &
O’Boyle, 2012).
Interestingly, in examining the measures that position engagement rela-
tive to burnout, all are dominated by questions about one’s affect or emo-
tion. Given that emotional exhaustion appears to be the most consistently
validated dimension of burnout across all burnout measurement instruments
(Schaufeli, Enzmann, & Girault, 1993), it is not surprising that engagement
measures heavily represented by questions about affect will show strong nega-
tive correlations with burnout. Further, it is important to note that despite the
multiplicity of published research studies in support of the burnout-antithesis
framework, Kahn fundamentally rejected the premise that engagement and
burnout are on the same continuum or opposites of each other, and he dis-
agreed with the premise that one cannot be simultaneously burned out and
engaged (Kahn, 2010). In fact, he described many examples he witnessed
during his field research of exhausted health care workers who met every
theory-based definition of being burned out, yet still had the capacity to
engage and regularly engaged completely in the treatment and care of their
patients (Kahn, personal interview, February 19, 2013). On this, he reflected
that engagement is not always about the energy an employee outwardly exhib-
its, but often about the energy an employee dedicates, visibly or not, to the
task(s) at hand (Kahn, 2010). In other words, an employee does not have
to be “bursting with energy” (Schaufeli et al., 2006, p. 714) in order to be
engaged (Shuck, 2011).
Job Satisfaction Framework. In the United States, the popular Gallup
Q12 is widely used as a well-understood benchmark for employee job sat-
isfaction and as a proxy for engagement (Harter et al., 2013). This defini-
tion of engagement, “an individual’s involvement and satisfaction with as
well as enthusiasm for work” (Harter, Schmidt, & Hayes, 2002, p. 269), was
developed from studies conducted with over 8,000 U.S. businesses in many

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


168 Anthony-McMann, Ellinger, Astakhova, Halbesleben

industries. Interestingly, the Q12 also includes a question that seeks to mea-
sure social connectedness at work (Harter et al., 2013). Robinson, Perryman,
and Hayday (2004) also developed a measure of engagement that operational-
ized a job-satisfaction–framed definition that positions engagement as a posi-
tive attitude toward an organization and its values. However, another study
showed that investing ‘one’s whole self,’ and not job satisfaction, correlated
more directly with engagement (Christian, Garza, & Slaughter, 2011). This
meta-analytic study also suggested that that job satisfaction (like burnout)
may be related to engagement, but it is not necessarily the same construct,
nor can it necessarily be measured by an instrument specifically designed to
measure job satisfaction (or burnout).
Multidimensional Frameworks. Alternatively, a number of multidimen-
sional models of employee engagement have emerged that highlight the com-
plexity of the construct by suggesting, for example, that job characteristics,
leadership actions, and personality traits are all antecedents to engagement
(Macey & Schneider, 2008). Drawing upon the work of many prior scholars,
Saks (2006) developed a multidimensional definition and model of employee
engagement that included cognitive, emotional, and behavioral components,
and he successfully tested this model against a number of antecedent and
outcome variables. This model specifically distinguished between job engage-
ment and organizational engagement suggesting that the former was linked
to an employee’s work-related role, and the latter was more closely tied to an
employee’s role within an organizational system.
A Return to Kahn. Despite the various engagement frameworks that
exist, many scholars are calling for a return to Kahn’s conceptualization
because it represents not just a broad motivational construct but also one with
specific psychological conditions that are both prerequisites and antecedents
(Saks & Gruman, 2014). Additionally, Kahn’s needs-satisfaction framework
implies a depth of consideration (i.e., the simultaneous investment of ener-
gies and the investment of one’s whole self) that seems inadequately served
by positioning it in relation to burnout or by measuring it through the lens of
job satisfaction. In the 20-plus years since he first presented his model, Kahn
(2010) has maintained his commitment to his needs-satisfaction framework,
rejecting those frameworks that place engagement in the shadow of burn-
out or job satisfaction. Indeed, despite the plethora of research that measures
engagement with the more narrowly framed UWES or Q12, many scholars
agree with him (Fletcher & Robinson, 2013; Saks & Gruman, 2014).
Workplace Stress
Early conceptualizations of workplace stress, which grew out of studies in the
biological and physical sciences, evolved significantly in response to the need
to understand the construct within the context of organizational settings. The
now widely accepted conservation of resources (COR) theory suggests that
people strive to keep and obtain valued resources (including objects, condi-

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


Exploring Different Operationalizations of Employee Engagement 169

tions, personal characteristics, and energies), and that they are threatened (or
stressed) by the loss, or threat of loss, of those resources (Hobfoll, 1989). The
findings from numerous studies of workplace stress reveal behavioral manifes-
tations that bear remarkable similarity to those described by Kahn (1990) as
stemming from disengaged employees. Although Hobfoll (1989) did not use
the term disengagement, he cited a familiar behavior pattern in people under
stress, stating that individuals experiencing stress will “strive to minimize net
loss of resources” (p. 517). In other words, they limit their engagement to the
extent it threatens their resources (Halbesleben, 2010). Positioned conversely,
this suggests that employees who are less stressed may also be more likely to
stay engaged.
Kahn (1990) contends that his three psychological preconditions of
engagement (meaningfulness, safety, and availability) are possible only when
an individual has his or her valued ‘resources.’ It follows, then, that the forces
that consume (or threaten) those resources contribute to stress and decrease
the likelihood of employee engagement. We examine these relationships with
two empirically sound engagement instruments: the Rich Scale (Rich et al.,
2010) and the ISA (intellectual, social, affective) Scale (Soane et al., 2012),
and a COR-based measure of workplace stress: the Workplace Stressors
Assessment Questionnaire (Mahmood, Coons, Guy, & Pelletier, 2010). Since
both engagement measures are based on Kahn’s (1990) needs-satisfaction con-
ceptualization, both should reveal negative relationships with workplace stress
(Figure 1).

HYPOTHESIS 1A: Workplace stress is negatively related to employee engagement as


measured by the Rich Scale.

Figure 1. Hypothesized Mediation Model

Burnout

H3a, H4a, H4b, H4c


H3b, H5a, H5b, H5c

Workplace Rich Employee


Stress H1a, H2 Engagement

H1b, H2 ISA Employee


Engagement

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


170 Anthony-McMann, Ellinger, Astakhova, Halbesleben

HYPOTHESIS 1B: Workplace stress is negatively related to employee engagement as


measured by the ISA Scale.

Given the apparent nomological differences between even similarly con-


ceptualized engagement measures, we are also suggesting that the relationship
between workplace stress and employee engagement differs depending on
the instrument used to measure employee engagement. Scholars concur with
Kahn (2007) that the presence of positive workplace relationships is a key
antecedent to the psychological conditions necessary under a needs-satisfac-
tion conceptualization of engagement (Schneider, Macey, Barbera, & Young,
2010). This suggests that the presence (or absence) of positive relationships
may impact the effects of workplace stress on employee engagement; however,
these relationships have not been studied. It also suggests that if the measure-
ment of engagement includes an element of interpersonal or social connected-
ness (i.e. positive relationships), the outcome of employee engagement may be
more resilient to those factors that seek to undermine it.
More specifically, we propose that the inclusion of the social (or inter-
personal relationship) dimension in the ISA Scale improves the resiliency of
engagement in the face of resource demands such as workplace stress and
therefore weakens the negative relationship between workplace stress and
engagement in comparison to the same relationship measured by the Rich
Scale.

HYPOTHESIS 2: The negative relationship between workplace stress and employee


engagement is stronger when employee engagement is measured by the Rich Scale
than when measured by the ISA Scale.

Burnout
The literature reveals relationships between engagement and burnout that are
complex—ranging from being conceptually distinct but related, to being con-
ceptually distinct but antipodean, to being at opposite ends of the same con-
ceptual continuum. However, studies of the relationships between workplace
stress and burnout are generally consistent. Initial studies on burnout focused
on the transactions and relationships between individuals at work (Maslach
et al., 2001). However, our understanding of burnout expanded significantly
in the early 1980s with the work of scholars who conceptualized burnout as
the consequence of prolonged exposure to stress (Maslach & Jackson, 1981).
Many years later, another specific link between stress and burnout was posited
by Gorgievski and Hobfoll (2008), who suggested, in accordance with COR
theory, that burnout was the unavoidable result of the chronic and steady
depletion of an individual’s resources or, in other words, through the buildup
of stress.

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


Exploring Different Operationalizations of Employee Engagement 171

Framed by exhaustion, cynicism, and reduced personal efficacy, burn-


out was operationalized by the widely used MBI (Maslach & Jackson, 1981).
Another broadly recognized conceptualization of burnout was offered by
Pines and Aronson (1988) as “a state of physical, emotional and mental
exhaustion caused by long-term involvement in situations that are emotion-
ally demanding” (p. 9). Other burnout instruments (such as the Oldenburg
Burnout Inventory) are similarly conceptualized, but they are also used to
measure disengagement—again, reflecting a burnout-antithesis–framed per-
spective of engagement (Demerouti & Bakker, 2008). Importantly, scholars
agree that the burnout instruments in broadest use in research and practice
reflect the underlying premise of burnout as resulting from the exhaustion
of an individual’s resources through prolonged exposure to stress (Schaufeli
et al., 1993).
Our next two predictions, Hypotheses 3a and 3b, propose that the pres-
ence of burnout (conceptualized in this study as analogous to exhaustion) has
a mediating or indirect effect on the relationship between workplace stress
and employee engagement, and that this indirect effect differs depending on
the instrument used to measure employee engagement. Although Kahn may
not have considered engagement as the opposite of or on a continuum with
burnout, research suggests that the two constructs may be related (Bakker,
Schaufeli, Leiter, & Taris, 2008). Consistent with the theoretical underpin-
nings of workplace stress, burnout, and Kahn’s (1990) conceptualization of
engagement, if workplace stress depletes valued resources that are needed
for the capacity for one to engage, and the accumulation of stress over time
can result in burnout, it is reasonable to predict that the eventual condi-
tion of burnout might explain the negative relationship between stress and
engagement. In other words, burnout has a mediating effect because it has a
meta capability that develops in individuals over time and after a prolonged
exposure to stress. Therefore, it is insufficient to suggest that workplace stress
alone has a negative influence on engagement.

HYPOTHESIS 3A: Burnout mediates the negative relationship between workplace stress
and employee engagement as measured by the Rich Scale.

HYPOTHESIS 3B: Burnout mediates the negative relationship between workplace stress
and employee engagement as measured by the ISA Scale.

Given some of the nomological differences in the Rich and ISA Scales,
the final set of hypotheses propose that the differences in the relationships
between workplace stress, burnout, and employee engagement can be found
at the dimensional level of the engagement operationalizations. Consistent
with the aforementioned description of the meta capability of burnout,
Hypotheses 4a, 4b, and 4c propose that burnout fully mediates the negative

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


172 Anthony-McMann, Ellinger, Astakhova, Halbesleben

relationships between workplace stress and the Rich dimensions of physical,


emotional, and cognitive engagement:

HYPOTHESIS 4A: Burnout fully mediates the negative relationship between workplace
stress and physical employee engagement as measured by the Rich subscale.

HYPOTHESIS 4B: Burnout fully mediates the negative relationship between workplace
stress and emotional employee engagement as measured by the Rich subscale.

HYPOTHESIS 4C: Burnout fully mediates the negative relationship between workplace
stress and cognitive employee engagement as measured by the Rich subscale.

Hypotheses 5a and 5c propose that burnout fully mediates the negative


relationships between workplace stress and the ISA dimensions of intellectual
and affective engagement. Again, it is expected that the meta-capabilities of
burnout in relation to workplace stress fully explain the negative relationships
between workplace stress and each of the above dimensions. On the other
hand, Hypothesis 5b proposes that burnout only partially mediates the nega-
tive relationship between workplace stress and ISA’s social engagement dimen-
sion. We suggest that the inclusion of the social (or interpersonal relationship)
dimension in the ISA Scale improves the resiliency of engagement (when mea-
sured by the ISA Scale). This mitigates the indirect effect of burnout on the
negative relationship between workplace stress and employee engagement
in comparison to the same relationship measured by the Rich Scale, which
excludes that dimension.

HYPOTHESIS 5A: Burnout fully mediates the negative relationship between workplace
stress and intellectual employee engagement as measured by the ISA subscale.

HYPOTHESIS 5B: Burnout partially mediates the negative relationship between work-
place stress and social employee engagement as measured by the ISA subscale.

HYPOTHESIS 5C: Burnout fully mediates the negative relationship between workplace
stress and affective employee engagement as measured by the ISA subscale.

Design and Method


The context for this survey-based study is the health care industry, a dynamic
environment that is often used in the study of employee engagement. During
the time of this study, a federal requirement to implement and ‘meaningfully
use’ electronic health records (EHRs) by 2015 required hospital information
technology (IT) professionals to work under significant stress while simulta-
neously remaining engaged to ensure successful implementations (Brooks &
Grotz, 2010). The measure of ‘meaningful use’ was specifically outlined by

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


Exploring Different Operationalizations of Employee Engagement 173

federal legislation and results were publicly reported. However, these EHRs
were relatively new, having only recently been engineered to function in the
resource-limited world of community hospitals, and there was immense pres-
sure to implement quickly in order to capitalize on federal incentive funds
as an offset to the significant capital outlays required for this technology
(DesRoches et al., 2010). Additionally, it was widely reported by health care
leaders that their IT employees found themselves increasingly responsible not
just for the veracity of the technology that underpinned the EHR but also for
direct support to clinicians during the patient care process.
Participants and Procedures
Specifically, IT professionals working on EHR implementations in community
hospitals in the United States were studied. This particular sample was chosen
because the IT professionals working to support EHR-related technologies
and processes represent a fairly homogenous group (Blumenthal & Tavenner,
2010). Regardless of the hospital for which they work, they were confronted
with similar technologies, clinical workflows, and objectives for implementa-
tion (Blumenthal & Tavenner, 2010). They were also facing the same time
pressures, working with the same complement of end users, juggling over-
whelming workloads, and experiencing many of the same stress factors. In
fact, empirical studies show that IT professionals often work in stress-charged
environments where burnout is not uncommon, yet they also exhibit high lev-
els of engagement in their work (Gan & Gan, 2013). We recruited participants
from a quota sample of hospitals that generally reflected the overall commu-
nity hospital population in terms of size and geographic location. Forty-five
community hospitals (or hospital systems), out of a total of 74 invited, par-
ticipated in the study based on confirmation, through conversations with chief
information officers (CIOs) or other senior IT executives, that EHR implemen-
tations were actively under way. An e-mail, drafted by the researchers, was
sent from the CIOs of the participating hospitals to all of their IT employees
with an invitation to participate in the study and a link to the web-based sur-
vey. A follow-up e-mail was sent two weeks later.
Of the 2,420 IT professionals recruited to participate, 472 completed
the survey. Of the respondents, 199 (42%) were female and 273 (58%) were
male, the mean age was 46 (SD = 10.08), and the mean organizational ten-
ure was 10 years (SD = 9.02). Twenty-seven (6%) respondents graduated from
high school, 100 (21%) attended college, 238 (50%) had an undergraduate
degree, and the remaining 107 (23%) started or completed a graduate degree.
Four hundred and nine (87%) respondents indicated that they work directly
with clients in some capacity, whereas the other 63 (13%) did not, and 123
(26%) respondents worked in some supervisory or managerial capacity. To test
for the possibility of nonresponse bias, a time trend analysis was conducted
in which survey respondents were split into two groups: those who com-
pleted the survey prior to the reminder e-mail that was sent two weeks after

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


174 Anthony-McMann, Ellinger, Astakhova, Halbesleben

the initial invitation to participate (early respondents), and those who com-
pleted the survey after the reminder e-mail was sent (late respondents). This
analysis assumes that late respondents are similar to nonrespondents in that
late respondents would have been nonrespondents had they not received the
reminder e-mail (Armstrong & Overton, 1977). The means for the two overall
engagement measures (Rich Scale and ISA Scale) were compared for the two
groups using independent samples t-tests, and Levene’s tests confirmed that
the variances between the two groups of respondents were equal (Hair, Black,
Babin, & Anderson, 2010). Further, both engagement measures evidenced
nonsignificant differences in the means between the two groups of respon-
dents (Rich Scale: t = −.152, p = .879; ISA Scale: t = 1.220, p = .223). These find-
ings support the assumption that the risk of nonresponse bias is low.
Measures
Participants were asked to rate each item from previously validated measure-
ment instruments on Likert scales ranging from ‘Strongly Disagree’ to ‘Strongly
Agree’ or ‘Never’ to ‘Always.’ Additionally, multiple-choice responses to certain
control variables, including gender, age, tenure, client interaction, and super-
visory status, were selected because previous studies in which engagement
was an outcome indicated that they might impact the hypothesized relation-
ships (Rich et al., 2010). Two other control variables, hospital size and EHR,
were also collected given the context of this study.
Workplace Stress. The 22-item Workplace Stressors Assessment Ques-
tionnaire (α = .95) assesses self-reported perceptions of workplace stress
(Mahmood et al., 2010). The choice of this instrument was determined by
its theoretical underpinning in COR and its psychometric development with
high-tech employees. Responses were scored on a 5-point Likert scale.
Burnout. The 10-item Burnout Measure, Short Version (BMS) (α = .85)
is consistent with this study’s exhaustion-based conceptualization of burn-
out in which burnout occurs after a prolonged exposure to stress (Malach-
Pines, 2005). Unlike other widely-used measures of burnout, this measure
has not also been used as a proxy for engagement, avoiding the possibility of
confounding the relationships between burnout and employee engagement.
Responses were scored on a 7-point Likert scale.
Employee Engagement. Two similarly conceptualized measures of
engagement were used in this study. The 18-item Rich Scale (α = .95) was
developed in congruence with Kahn’s (1990) needs-satisfaction framework
and tested with 245 firefighters and their supervisors (Rich et al., 2010). This
scale is organized into three subscales or dimensions (physical, emotional, and
cognitive), and all responses were scored on a 5-point Likert scale. In order to
explore the nomological framework of this scale, responses were evaluated at
the overall level (all 18 questions) and at the subscale level.
The second engagement scale, the nine-item ISA Scale ( α = .91)
(Soane et al., 2012), was developed and tested among 540 employees in a

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


Exploring Different Operationalizations of Employee Engagement 175

United Kingdom (UK)-based manufacturing company and, like the Rich


Scale, is underpinned by the work of Kahn (1990). The scale is also orga-
nized into three subscales, or dimensions (intellectual, affective, and social),
and responses were scored on a 7-point Likert scale. Of particular interest,
the social dimension appears to reflect the element of positive interpersonal
relationships that Kahn (1990) saw as fundamental to an employee’s willing-
ness to engage at work. Responses were also evaluated at the overall level (all
nine questions) and at the subscale level.

Results
We begin with a review of the preliminary analyses, including confirmatory
factor analyses, discriminant validity, interclass correlation coefficients, corre-
lations, and descriptive statistics. Next, we present the results of our hypoth-
esis testing using multiple hierarchical regression and maximum likelihood
structural equation modeling. All models were evaluated in accordance with
the recommended combination of fit indices (Hu & Bentler, 1999). The rela-
tionships were tested by first verifying the significance of the direct effects
(between workplace stress and the various measures/dimensions of engage-
ment), and then, we added the mediating variable of burnout and verified
the significance of the relationships. When the presence of mediation was
detected, the magnitude of the indirect effects were analyzed via bootstrap-
ping techniques at a 95% confidence interval using 5,000 bootstrapped sam-
ples (Preacher & Hayes, 2004). To determine if an indirect effect was full or
partial, we compared the models with all the variables to those in which the
direct effect was constrained (Baron & Kenny, 1996).
Preliminary Analyses
First, we verified the assumptions required for multivariate analysis. Then,
through confirmatory factor analysis (CFA), using maximum likelihood estima-
tion, we confirmed the discriminant validity of each latent variable used in the
study (Hair et al., 2010). To rule out the possibility of bias due to common
method variance, both the Harmon single-factor test (using exploratory factor
analysis) and the common latent factor test (using CFA) were used (Podsakoff,
MacKenzie, Jeong-Yeon, & Podsakoff, 2003). Because survey participants were
nested in different hospital organizations, another test controlled for the possibil-
ity of bias due to multilevel effects. Interclass correlation coefficients (ICC1 and
ICC2) were calculated for each construct (and each subscale or dimension) and
for each hospital or hospital system that participated in the study (McGraw &
Wong, 1996). The calculated ICCs were all nonsignificant [ICC1s ranged from
.022 to .335, (p > .05); ICC2s ranged from .011 to .322, (p > .05)], indicating that
the potential for interorganizational bias was not likely (Shrout & Fleiss, 1979).
A summary of means, standard deviations, internal reliabilities, and
zero-order correlations is presented in Tables 1 and 2 . The Cronbach ’s

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


176 Anthony-McMann, Ellinger, Astakhova, Halbesleben

Table 1. Descriptive Statistics, Reliabilities and Zero-Order Correlation


Coefficients of and Among the Study Variables With Engagement
Measured by the Rich Scale
M SD AVE CR 1 2 3 4 5 6
1. Workplace 2.63 .67 84.54 .99 .92
Stress
2. Burnout 3.01 1.10 84.10 .98 .71** .92
3. Physical 4.54 .52 57.74 .89 –.05 –.05 .88
Engagement
4. Emotional 4.15 .74 54.05 .92 –.48** –.55** .50** .92
Engagement
5. Cognitive 4.34 .60 70.46 .93 –.10* –.10* .73** .50** .93
Engagement
6. Overall Rich 4.34 .52 87.22 .99 –.28** –.31** .85** .83** .86** .94
Engagement
Note: n = 472; M = mean; SD = standard deviation; AVE = percentage of average variance extracted;
CR = composite reliability; Cronbach’s alphas (α) are presented diagonally.
*p < .05; **p < .01.

Table 2. Descriptive Statistics, Reliabilities and Zero-Order Correlation


Coefficients of and Among the Study Variables with Engagement
Measured by the Intellectual, Social, Affective (ISA) Scale
M SD AVE CR 1 2 3 4 5 6
1. Workplace 2.63 .67 84.64 .99 .92
Stress
2. Burnout 3.01 1.10 84.10 .98 .71** .92
3. Intellectual 6.36 .78 85.26 .95 –.17** –.18** .94
Engagement
4. Social 4.96 1.35 81.05 .93 –.43** –.34** .18** .93
Engagement
5. Affective 5.80 1.18 78.20 .91 –.44** –.53** .52** .42** .91
Engagement
6. ISA Overall 5.71 .85 87.11 .98 –.49** –.48** .64** .78** .85** .87
Engagement
Note: n = 472; M = mean; SD = standard deviation; AVE = percentage of average variance extracted;
CR = composite reliability; Cronbach’s alphas (α) are presented diagonally.
**p < .01.

alphas for the scales are acceptable, ranging from .87 to .94. In support
of the internal and convergent validity of the constructs, the percentages
of average variance extracted are also acceptable (ranging from 54.05% to
87.22%), as are the composite reliabilities (ranging from .89 to .99) (Hair

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


Exploring Different Operationalizations of Employee Engagement 177

et al., 2010). With the exception of the nonsignificant correlations between


the Rich subscale of physical engagement and both workplace stress and
burnout, all other correlations are significant and within acceptable ranges.
As expected, the correlations between both measures of overall engagement
and each of their subscales are relatively high (r ranging from .64 to .86);
however, the social dimension of the ISA Scale had lower correlations to
the other ISA subscales (ranging from .18 to .42) than might be expected in
a single construct. The bivariate correlation between workplace stress and
burnout is relatively high (r = .71), but this is consistent with the literature,
which suggests that stress and burnout are highly interrelated (Gorgievski
& Hobfoll, 2008).
Hypothesis Tests
As our intent was to deconstruct the two measures of employee engagement
in order to understand the differences in their underlying nomological frame-
works, four measurement models were developed for this study—two with
the overall measures of engagement and two with dimensional specificity
(Table 3). The impact of the control variables showed no significant impact
on the hypothesized relationships, so, for the sake of parsimony, the models
and their fit indices are shown here without them. The discriminant valid-
ity of the constructs was confirmed by comparing each of the four measure-
ment models to two alternative models (Hair et al., 2010). The first alternative
model combined the items of workplace stress and burnout into one fac-
tor in consideration of the high interrelations between those two constructs
(Gorgievski & Hobfoll, 2008). The second alternative model combined all
of the items from all of the constructs into a single factor. As shown in Table
3, the combination of fit indices for the hypothesized measurement models
all indicate a better fit than those of their alternative models (Hu & Bentler,
1999). These results support the discriminant validity for each construct as
well as the discriminant validity of the three engagement dimensions in each
of the engagement measures (Hair et al., 2010). This is an important finding
because the construct validity of employee engagement has been the subject
of much debate (Christian et al., 2011).
The results of the hypotheses testing are depicted in Figures 2 and 3. As
predicted by Hypotheses 1a and 1b, there are significant correlations between
workplace stress and Rich overall engagement (r = −.28, p < .01) and between
workplace stress and ISA overall engagement (r = −.49, p < .01). Hypoth-
esis 2, which predicted a stronger negative relationship between workplace
stress and engagement when engagement is measured by the Rich Scale, is
not specifically supported. But this hypothesis also suggests that there is a
significant difference in the relationships between workplace stress and the
two overall measures of employee engagement. Fisher’s r to z calculation con-
firms that these relationships are, in fact, significantly different (z = 3.804,
p < .001) (Preacher, 2002). Further, although not specifically hypothesized,

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


178 Anthony-McMann, Ellinger, Astakhova, Halbesleben

Table 3. Measurement Models and Confirmatory Factor Analysis (CFA)


Model/Fit Indices df χ2 Δχ2(df) NNFI CFI RMSEA SRMR

A. Rich Overall, 148 770.91* .95 .96 .09 .07


Workplace Stress,
Burnout
Alternative Model A1a 150 1,169.52* 398.61(2)* .92 .93 .12 .08
Alternative Model A2b 151 1,358.04* 587.13(3)* .91 .92 .13 .09
B. ISA Overall, 147 743.60* .96 .96 .09 .08
Workplace Stress,
Burnout
Alternative Model B1c 149 1,146.18* 402.49(2)* .93 .94 .12 .09
Alternative Model B2b 150 1,279.85* 536.16(3)* .92 .93 .13 .08
C. Rich Subscales, 517 1,797.48* .96 .97 .07 .07
Workplace Stress,
Burnout
Alternative Model C1d 521 2,198.10* 400.62(4)* .95 .96 .08 .08
Alternative Model C2b 527 7,271.95* 5,474.5(10)* .81 .82 .16 .20
D. ISA Subscales, 265 1,010.65* .96 .97 .08 .06
Workplace Stress,
Burnout
Alternative Model D1e 269 1,442.18* 432.16(4)* .94 .95 .10 .08
Alternative Model D2b 275 4,543.50* 3,532.9(10)* .80 .81 .18 .14
Note: df = degrees of freedom; χ2 = chi-square; Δχ2 = change
in chi-square; NNFI = non-normed
fit index; CFI = comparative fit index; RMSEA = root mean square error of approximation;
SRMR = standardized root mean square residual; ISA = intellectual, social, affective.
aAlternative Model A1 loads on two factors: Rich overall engagement and the combined items of
workplace stress and burnout.
bAlternative Models A2, B2, C2 and D2 load all items on a single factor.

cAlternative Model B1 loads on two factors: ISA overall engagement and the combined items of
workplace stress and burnout.
dAlternative Model C1 loads on four factors: physical engagement, emotional engagement, cognitive
engagement, and the combined items of workplace stress and burnout.
eAlternative Model D1 loads on four factors: intellectual engagement, social engagement, affective
engagement, and the combined items of workplace stress and burnout.
*p < .001.

the relationships between burnout and the two overall measures of employee
engagement are also significantly different (z = 3.1, p < .01).
Hypotheses 3a and 3b proposed that burnout explains the negative rela-
tionships between workplace stress and employee engagement as measured
by the Rich Scale and the ISA Scale, respectively. The combination of model
fit indices are acceptable for the Rich overall scale [ χ 2(148) = 770.91;

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


Exploring Different Operationalizations of Employee Engagement 179

Figure 2. Mediation Models With Rich Overall Scale and ISA Overall Scale

Burnout

.75***/.75*** -.28*/-.47***

Overall
Workplace
Employee
Stress -.12*(-.38**)/ Engagement
-.14*(-.58***)

Standardized path coefficients (SPCs) are depicted, and the SPC in parentheses represents the
coefficient from the base model. The first coefficient relates to the Rich overall employee engagement,
and the second coefficient relates to the intellectual, social, affective (ISA) overall employee
engagement.
*p < .05; **p < .01; ***p < .001.

non-normed fit index (NNFI) = .95; comparative fit index (CFI) = .96; root
mean square error of approximation (RMSEA) = .09; standardized root mean
square residual (SRMR) = .07] and for the ISA overall scale [χ2(147) = 743.69;
NNFI = .96; CFI = .96; RMSEA = .09; SRMR = .08]. The standardized path
coefficients (SPCs) in Figure 2 confirm the significant relationships between
workplace stress and both overall measures of engagement, between work-
place stress and burnout, and between burnout and both overall engagement
measures. Further, as Figure 2 shows, the relationships between workplace
stress and both overall engagement measures are significant and smaller (Rich
overall: SPC = −.12, p < .05; ISA overall: SPC = −.14, p < .05) than they are in
base models which exclude the burnout variable (Rich base model: SPC = −.38,
p < .01; ISA base model: SPC = −.58, p < .001). Burnout’s indirect effect on the
relationships between workplace stress and both overall engagement measures
was confirmed by bootstrapping analysis (Rich overall: CI = −.20; −.06; ISA
overall: CI = −.36; −.14).
Finally, partial mediation (as opposed to full mediation) was confirmed
by comparing each full model to an alternative model in which the direct
path (between workplace stress and engagement) was constrained. The Rich
Scale full model has a lower chi-square than its alternative model and the dif-
ference in the chi-squares is significant [Δχ2(1) = 10.24, p < .05]. The differ-
ences between the ISA Scale full model and its alternative are not significant
[Δχ2(1) = 3.77, p > .05], which implies that burnout may fully mediate the
negative relationship between workplace stress and ISA overall engagement.

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


180 Anthony-McMann, Ellinger, Astakhova, Halbesleben

Figure 3. Mediation Models With Rich and ISA Subscales

Burnout

.00/-.09
. 75***/.75*** Cognitive/Intellectual
Employee
Engagement

-.40***/-.43***

-.22**/-.19*
Emotional/Affective
Workplace Employee
Stress -.28**(-.62***)/ Engagement
-.20**(-.56***)
.02/.01
-.18*/-.51***
Physical/Social
Employee
Engagement

Standardized path coefficients (SPCs) are depicted, and the SPC in parentheses represents the
coefficient from the base model. The first coefficient relates to the Rich dimensional levels of
employee engagement, and the second coefficient relates to the intellectual, social, affective (ISA)
dimensional levels of employee engagement.
*p < .05; **p < .01; ***p < .001.

However since the partially mediated model displays significant relationships


among all the variables and since our bootstrapping analysis confirms these
relationships, Hypothesis 3b is supported.
Hypotheses 4a, 4b, and 4c proposed that burnout will fully explain the
negative relationships between workplace stress and each of the three Rich
subscales: physical, emotional and cognitive engagement. The combination
of fit indices are acceptable for the Rich subscale model depicted in Figure 3
[χ2(520) = 2308.17; NNFI = .95; CFI = .95; RMSEA = .09; SRMR = .17]. There are
significant relationships between workplace stress and both physical and cogni-
tive engagement and between workplace stress and burnout. However, the rela-
tionships between burnout and physical engagement and between burnout and
cognitive engagement are nonsignificant, indicating that burnout has no indi-
rect effect on those dimensions (Hair et al., 2010). Turning to emotional engage-
ment, there are significant relationships among all three constructs. As Figure 3
shows, the relationship between workplace stress and emotional engagement

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


Exploring Different Operationalizations of Employee Engagement 181

Table 4. Structural Models


Model/Fit Indices df χ2 Δχ2(df) NNFI CFI RMSEA SRMR

A. Rich Overall Base 149 827.17* .95 .96 .10 .07


Modela
Rich Overall Full Modelb 148 770.91* 56.26(1)* .95 .96 .09 .07
B. ISA Overall Base 148 787.57* .95 .96 .10 .08
Modelc
ISA Overall Full 147 743.69* 43.88(1)* .96 .96 .09 .08
Modeld
C. Rich Subscales Base 523 2,344.33* .95 .95 .09 .17
Modele
Rich Subscales Full 520 2,308.17* 36.16(3)* .95 .95 .09 .17
Modelf
D. ISA Subscales Base 271 1,218.61* .95 .96 .09 .09
Modelg
ISA Subscales Full 268 1,182.72* 35.89(3)* .96 .96 .09 .09
Modelh
Note: df = degrees of freedom; χ2 = chi-square; Δχ2 = change in chi-square; NNFI = non-normed
fit index; CFI = comparative fit index; RMSEA = root mean square error of approximation;
SRMR = standardized root mean square residual; ISA = intellectual, social, affective.
a Base Model A includes Rich overall engagement and workplace stress.

b Full Model A includes Rich overall engagement, workplace stress, and burnout.

c Base Model B includes ISA overall engagement and workplace stress.

d Full Model B includes ISA overall engagement, workplace stress, and burnout.

e Base Model C includes physical engagement, emotional engagement, cognitive engagement, and
workplace stress.
f Full Model C includes physical engagement, emotional engagement, cognitive engagement,
workplace stress, and burnout.
g Base Model D includes intellectual engagement, social engagement, affective engagement, and
workplace stress.
h Full Model D includes intellectual engagement, social engagement, affective engagement,
workplace stress, and burnout.
*p < .001.

is significant and smaller (SPC = −.28, p < .01) than it is in a base model, which
excludes the burnout variable (SPC = −.62, p < .001). Bootstrapping analysis
supports the significance of the mediating relationship and resultant indirect
effect (CI = −.42; −.24). Comparisons to an alternative model in which the direct
relationship is constrained confirm Hypothesis 4b in that burnout’s indirect
effect on emotional engagement is partial [Δχ2(1) = 15.73, p < .001].
Finally, Hypotheses 5a and 5c proposed that burnout will fully explain
the negative relationships between workplace stress and two of the three ISA
subscales: intellectual and affective engagement. Hypothesis 5b proposed that
burnout will partially explain the negative relationship between workplace

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


182 Anthony-McMann, Ellinger, Astakhova, Halbesleben

stress and ISA social engagement. The combination of fit indices are accept-
able for the ISA subscale model depicted in Figure 3 [χ2(268) = 1182.72;
NNFI = .96; CFI = .96; RMSEA = .09; SRMR = .09] (Hu & Bentler, 1999). Start-
ing with intellectual and social engagement, there are significant relationships
between workplace stress and both intellectual and social engagement and
between workplace stress and burnout. However, the relationships between
burnout and intellectual engagement and between burnout and social engage-
ment are nonsignificant, indicating that burnout has no indirect effect on
those dimensions. Concluding with affective engagement, there are significant
relationships among all three constructs. As Figure 3 shows, the relationship
between workplace stress and affective engagement is significant and smaller
(SPC = −.20; p < .01) than it is in a base model, which excludes the burnout
variable (SPC = −.56, p < .001). Bootstrapping analysis supports the signifi-
cance of the mediating relationship and resultant indirect effect (CI = −.70;–
.38). Comparisons to an alternative model in which the direct relationship is
constrained confirm Hypothesis 5c in that burnout’s indirect effect on affective
engagement is partial [Δχ2(1) = 7.85, p < .05].

Discussion
A recent research report, “Employee Job Satisfaction and Engagement: Opti-
mizing Organizational Culture for Success,” published by the Society for
Human Resource Management (SHRM, 2015), underscores the importance
of maintaining an engaged workforce and understanding the conditions that
promote and improve engagement. However, to improve engagement, orga-
nizational leaders and managers need to measure it, and the use of the appro-
priate engagement measurement instrument is a critical first step. Given the
abundance of engagement measures, practitioners and researchers face a dif-
ficult choice. Without truly understanding the differences among the instru-
ments, incorrect operationalizations of the construct may occur, resulting in
conceptual and measurement fallacies.
Our study attempts to clarify two such measures—both of which are
theoretically underpinned by the same needs-satisfaction framework. As
predicted, these two employee engagement measures (analyzed at both the
overall and dimensional levels) show significantly different relationships with
workplace stress and burnout. A summary of the overall findings associated
with each hypothesis is presented in Table 5. However, a review of the empiri-
cal findings within the context of each measure’s conceptual framework offers
further clarity.
First, both needs-satisfaction–based measures of employee engagement
examined in this study reveal the predicted negative relationships with work-
place stress. This finding is consistent with Hobfoll’s (1989) conservation of
resources theory of stress (in that stress results when valued resources like
time or positive working conditions are lost) and Kahn’s (1990) theoreti-

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


Exploring Different Operationalizations of Employee Engagement 183

Table 5. Comparison of Employee Engagement Measures and Their


Relationships to Workplace Stress and Burnout
Negative Relationship With Negative Relationship
Engagement Measure Workplace Stress With Burnout
Rich Overall Engagement Moderate/Partially mediated Weak (H3a)
by burnout (H1a, H2, H3a)
• Physicala Nonsignificant (H4a) Nonsignificant (H4a)
• Emotionalb Strong/Partially mediated by Strong (H4b)
burnout (H4b)
• Cognitivec Weak (H4c) Nonsignificant (H4c)

ISA Overall Engagement Strong/Partially mediated by Strong (H3b)


burnout (H1b, H2, H3b)
• Intellectualc Weak (H5a) Nonsignificant (H5a)
• Sociala Strong (H5b) Nonsignificant (H5b)
• Affectiveb Strong/Partially mediated by Strong (H5c)
burnout (H5c)
a Unique (dissimilar) dimensions.
b Conceptually and empirically similar dimensions.
c Conceptually and empirically similar dimensions.

cal conceptualization of engagement (as engagement requires psychological


preconditions which, in turn, require valued resources). Next, we see that
the two employee engagement measures (and their subscales) reveal signifi-
cantly different relationships with workplace stress and burnout. As these two
engagement measures are similarly conceptualized based on Kahn’s (1990)
needs-satisfaction framework, it is useful to examine their similarities first.
Similarities Between the Rich and ISA Engagement Measures
The Rich Scale (Rich et al., 2010) and ISA Scale (Soane et al., 2012) both
incorporate three subscales or dimensions of employee engagement. The Rich
dimensions are physical, emotional, and cognitive, and the ISA dimensions
are intellectual, social, and affective. It is not surprising, because both mea-
sures are grounded in Kahn’s (1990) conceptualization, that some of their
dimensions (or subscales) are very similar to each other, and these similarities
reflect the existing research on employee engagement. For example, the Rich
cognitive subscale and ISA intellectual subscale demonstrate both conceptual
and empirical similarities. Rich et al. (2010) describe cognitive engagement as
both a measure of “attention (level of amount of focus and concentration) and
absorption (level of engrossment or the intensity of the focus and concentra-
tion)” (pp. 623–624). Similarly, the ISA intellectual dimension is described
as “the extent to which one is intellectually absorbed in work” (Soane et al.,
2012, p. 532). To further support that these subscales are measuring similar

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


184 Anthony-McMann, Ellinger, Astakhova, Halbesleben

aspects of employee engagement, the empirical data from this study show that
these two subscales both show weak (but almost identical) negative relation-
ships with workplace stress and no significant relationship with burnout.
This suggests that although, as predicted, an employee’s focus or absorp-
tion at work may be hindered by workplace stress, it is not impacted by the
emotional exhaustion reflected by burnout. Considered in the context of con-
servation of resources theory (Hobfoll, 1989), the resource loss associated
with workplace stress and burnout appears not to be strong enough to materi-
ally influence an employee’s ability to concentrate. It is possible that the nature
of IT work, which generally requires a high degree of technical precision and
focus, may attract the type of employee who can remain ‘in role’ despite the
resource loss associated with workplace stress and/or burnout (Kahn, 1992).
Nevertheless, workplace stress and burnout display almost identical results
with the outcomes of these dimensions in both the Rich and ISA scales.
Similarly, the Rich emotional subscale and the ISA affective subscale also
demonstrate conceptual and empirical similarities. Rich et al. (2010) describe
emotional engagement as a measure of “enthusiasm, happiness and optimism
experienced at work” (p. 623) and is based on research about core affect
(Russell & Barrett, 1999). Similarly, the ISA affective dimension is described as
“the extent to which one experiences a state of positive affect [emotion] relat-
ing to one’s work role” (Soane et al., 2012, p. 532). Our study supports the
conclusion that the Rich emotional engagement and the ISA affective engage-
ment subscales are measuring similar aspects of engagement because both
show significant negative relationships with workplace stress and significant
negative relationships with burnout. Further, burnout partially explains the
negative relationships with workplace stress and both engagement subscales.
Even though it was predicted that burnout would fully mediate these relation-
ships, this finding is still consistent in that an employee’s emotional state is
highly impacted by the resource loss associated with both workplace stress
and burnout (Hobfoll, 1989). As this study’s conceptualization of burnout is
analogous to emotional exhaustion, its negative impact on emotional/affective
engagement in the Rich and ISA scales makes sense.
Given the above similarities, the two remaining dimensions (i.e.,
the physical engagement from the Rich Scale and social engagement from
the ISA Scale) must explain the significant differences in the relationships
when engagement is measured at the overall (all subscales) level. Indeed,
the remaining two dimensions measure fundamentally different aspects of
employee engagement and, in this study, both reveal very different relation-
ships with workplace stress and burnout.
Differences Between the Rich and ISA Engagement Measures
The Rich physical dimension was the only dimension (in either engagement
measure) that shows no significant correlation with either workplace stress
or burnout (Table 1). Further, the weak relationship with workplace stress

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


Exploring Different Operationalizations of Employee Engagement 185

and physical engagement (Figure 3) is significant only in the presence of the


other subscales. Rich et al. (2010) describe physical engagement as based
upon a combination of work intensity as conceptualized by Brown and Leigh
(1996) and physical engagement as conceptualized by Kahn (1992). This six-
item subscale includes questions about “working with intensity,” “exerting full
effort,” “trying [one’s] hardest,” and “exerting a lot of energy on [the] job”
(Rich et al., 2010, p. 634).
The nonsignificant findings in this study suggest that neither workplace
stress nor burnout affects an employee’s physical engagement or level of
energy/intensity at work. Seemingly counterintuitive, there may be reason-
able explanations for this. First, the specific context of this study—namely,
IT professionals working to implement EHRs in U.S. hospitals—might skew
the relationship between workplace stress and physical engagement due to
the unique nature of the circumstances under which the survey respondents
were working. All of these professionals were working toward a federally man-
dated deadline for EHR adoption, and the consequences for not meeting that
deadline included significant financial penalties for their hospital employers.
As such, it is possible that the energy they exhibited towards their work may
have been more a function of the pressure they felt than the sense of physical
engagement that the measure was intending to capture.
Another explanation may reside within the physical engagement survey
questions themselves. Unlike the other questions that relate to concentration,
focus, happiness, or social connectedness, the physical dimension is the only
one in which respondents indicated their perceptions of how hard they were
working. Although attempts to minimize the possibility of bias due to social
desirability (Tsai & Ghoshal, 1998) were made, there is still a possibility that
this particular set of questions, which asks respondents about perceptions of
their own personal work ethic, was influenced either by cognitive dissonance
or by a reluctance to admit something thought to be inconsistent with their
choice to remain in their jobs (Bertrand & Mullainathan, 2001). This cogni-
tive effect could result in both high mean scores (which we see in Table 1) and
a confounded (or, as in this study, nonsignificant) relationship with workplace
stress.
Turning to the ISA social engagement subscale, this dimension is also
unique because it shows the predicted strong negative relationship with work-
place stress, but no relationship with burnout. Conceptually, this three-item
dimension appears to measure an aspect of engagement not contemplated by
the Rich Scale. It is described as “the extent to which one is socially connected
with the working environment and shares common values with colleagues”
(Soane et al., 2012, p. 532). This dimension is grounded in Kahn’s clear con-
tention that connectedness or positive interpersonal relationships at work are
key to employee engagement (Kahn, 2007). From an empirical perspective,
Table 2 shows that the mean for social engagement is well above the midpoint
of its scale (4.96 out of 7.00). This suggests that the level of social connected-

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


186 Anthony-McMann, Ellinger, Astakhova, Halbesleben

ness or positive interpersonal relationships at work (as measured by the ISA


social dimension) is fairly strong in this sample of IT professionals despite its
strong negative relationship with workplace stress. The mediating effect of
burnout was hypothesized to only partially explain the negative relationship
between workplace stress and social engagement based on the contention that
positive interpersonal relationships at work improve the resiliency of engage-
ment in relation to those variables that may seek to negatively influence it.
However, it appears that these relationships, with both the working environ-
ment and among work colleagues, are strong enough (even in the face of
workplace stress) to completely resist the negative effect of burnout. These
findings support Kahn’s (2007) position about the positive effect of strong
interpersonal relationships on employee engagement and are consistent with
Hobfoll’s (1989) premise that these relationships are resources that work to
offset the negative effects of stress or burnout.
Impact of Subscales on Overall Measures of Employee Engagement
The Rich overall engagement scale shows fairly weak relationships with both
workplace stress and burnout. This is likely due to the very weak relation-
ship between both independent variables and its cognitive dimension and the
nonsignificant relationship between both independent variables and its physi-
cal dimension. On the other hand, the ISA overall engagement scale shows
fairly strong negative relationships with the two independent variables. The
strength of the relationship with workplace stress is likely due to the strong
relationship with both affective and social engagement, and the relationship
with burnout is likely due to the strong relationship with its affective dimen-
sion. Yet in the mediated models with burnout and the overall measures of
employee engagement (Figure 2), the resultant relationships between both
overall measures and workplace stress are almost identical. In other words,
burnout explains more of the negative relationship between workplace stress
and the ISA Scale than it explains between workplace stress and the Rich
Scale. This suggests that the ISA overall measure is more sensitive to the effect
of burnout despite the fact that the correlation analysis indicates that two of
its three dimensions have nonsignificant relationships with burnout. Again,
an explanation for this is found when the relationships with the ISA subscales
and independent variables are modeled simultaneously.
As discussed, affective engagement has a strong negative relationship
with burnout, and although intellectual engagement has a nonsignificant rela-
tionship with burnout on its own, in the presence of the other ISA subscales,
it also has a significant, albeit weak, negative relationship with burnout. This
suggests that the negative effect of burnout on the combination of affective
and intellectual engagement serves to increase both the negative relationship
between burnout and ISA overall engagement and the indirect effect of burn-
out on the ISA overall engagement scale in comparison to the Rich overall
engagement scale. However, social engagement reflects interconnectedness,

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


Exploring Different Operationalizations of Employee Engagement 187

positive shared values, and positive relationships among coworkers (Soane


et al., 2012); consistent with the literature, the presence of this phenomenon
appears to weaken the negative effect of burnout on social engagement. Were
it not for an examination of the ISA subscales, the positive impact of social
engagement (and the interventions it suggests) could be overshadowed by the
negative relationships evidenced with the other two ISA dimensions.

Employee Engagement and Burnout


Although not specifically hypothesized, we also examined the relationship
between burnout and employee engagement in the hope of addressing some
of the tensions in the literature about these phenomena (Fletcher & Robin-
son, 2013; Saks & Gruman, 2014). Despite the preponderance of research
that positions burnout either on the same continuum as engagement or as the
antipode of engagement, burnout exhibits either very weak or no relationship
with any dimension of employee engagement that is not measuring emotion.
Although the findings show significant negative relationships between burn-
out and both overall measures of employee engagement, an examination of
the measures’ subscales reveal that this negative relationship is almost entirely
attributable to the emotional/affective dimensions. This suggests that engage-
ment measures that are conceptually dominated by the concepts of affect or
emotion (e.g., the MBI, UWES, and OLBI) may, in all likelihood, be measuring
the antithesis of burnout. Nevertheless according to Kahn and an increasing
number of scholars, the antithesis of burnout is not the same as employee
engagement (Saks & Gruman, 2014).

Implications for Theory


A consistent understanding of employee engagement remains elusive due to
its many conceptualizations and operationalizations (Keenoy, 2014; Shuck
et al., 2012). Therefore, one of this study’s most significant contributions to
scholarly literature is that it offers a nuanced explanation of some of the dif-
ferences in how employee engagement is measured and how those differences
may impact human resource development (HRD) efforts. Although this study
employed two needs-satisfaction–based measures of employee engagement to
illustrate this point, its dimensional-level findings may also clarify measures
of engagement that are conceptualized based on a burnout-antithesis frame-
work. Specifically, the findings with the Rich emotional dimension and the
ISA affective dimension, which show strong negative relationships with burn-
out, suggest that burnout-antithesis–framed measures of employee engage-
ment (which are largely dominated with questions about affect) are, in fact,
measuring the antithesis of burnout or, said another way, the antithesis of
emotional exhaustion. Although scholars may debate whether the antithesis of
burnout is, in fact, the same thing as employee engagement, this study offers
additional insight into what, exactly, these measures may be measuring.

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


188 Anthony-McMann, Ellinger, Astakhova, Halbesleben

The key premise of this study was that even similarly conceptualized
measures of employee engagement would reveal different relationships with
workplace stress and burnout, and that these differences would be impor-
tant in understanding what the measures are actually measuring. Our study
strongly suggests that, regardless of conceptualization, employee engage-
ment is domain specific, and thus the meaning of the construct is revealed
only upon examination of the dimensional level of engagement instruments.
Indeed, this study confirms that although both measures of employee engage-
ment employed in the proposed study are conceptually based on Kahn’s
(1990) needs-satisfaction framework, even they do not measure the same
aspects of employee engagement. As a result, different suggestions for the util-
ity of each measure and different workplace interventions suggested by each
measure can be teased out of these findings.
Finally, much of the recent research on employee engagement has been
based on burnout-antithesis–framed measures of engagement (Saks & Gru-
man, 2014). Although burnout is certainly an interesting lens through which
to evaluate employee engagement, it is, by definition, a narrow one. There
is little doubt that Kahn’s (1990) conceptualization has been challenging
to operationalize; however as scholars return to his model, they will likely
embrace an understanding of employee engagement that is more holistically
consistent with desired workplace outcomes—at both employee and orga-
nizational levels. In other words, it seems likely that the breadth and depth
of outcome improvements would expand when a workforce is engaged at
intellectual, cognitive, physical, and social levels, in addition to affective or
emotional ones.
Implications for Practice
It can be reasonably inferred from this study that different employee engage-
ment instruments (even those that are similarly conceptualized) will all have
different predictive properties and thus different utility. However, if organiza-
tional leaders and managers want to obtain a broad picture of the engagement
levels of their employees, then the choice of employee engagement instrument
may be less critical. For example, since all of the studied engagement dimen-
sions are predicated on the presence of enthusiasm, optimism, and positive
emotions at work, workplace interventions designed to mitigate the nega-
tive effects of stress and/or lessen the likelihood of emotional exhaustion may
serve to both curtail the development of burnout and improve the resilience
of employee engagement. However if the goal is to better understand why
stress translates into lower engagement levels, then the choice of measurement
instrument may determine the outcome. For example, if organizational lead-
ers are interested in measuring the impact of workplace stress on the affective
or emotional elements of engagement, then either a burnout-antithesis–based
measure of engagement or the emotional/affective dimensions from the Rich
Scale or ISA Scale may be appropriate. Alternatively, if organizations seek to

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


Exploring Different Operationalizations of Employee Engagement 189

understand the impact of stress on interpersonal relationships or social con-


nectedness at work, then ISA’s social dimension may be the only measure
sensitive to that specific aspect of employee engagement.
The selection of employee engagement measurement instruments should
also be evaluated in the context of what the organization is considering in
terms of workplace interventions (Shuck, 2011). This study suggests that dif-
ferent dimensions of engagement may be more or less sensitive to the impact
of certain interventions. For example, the Rich Scale findings suggest that
since neither workplace stress nor burnout have a strong negative effect on
engagement, then workplace environments that are generally meaningful to
employees, psychologically safe, and supported by sufficient resources will
likely result in an engaged workforce (Kahn, 1990). On the other hand,
the ISA Scale findings, which reveal much stronger relationships with both
workplace stress and burnout, also show that social connectedness or strong
interpersonal relationships at work serve to mitigate the negative effects of
both. This implies that while stress and burnout do, in fact, impact employee
engagement, efforts to build a positive, trusting culture may be a powerful
force against those negative effects. Further, these findings lend credence to
Kahn’s (2007) contention that the most important engagement-related initia-
tive an organization can undertake might be one in which leaders, coworkers
and peers are trained on how to create and foster environments conducive
to the development of positive relationships at work. However, an interven-
tion suggested by the ISA Scale (e.g., the launch of a leadership development
program focused on training leaders to create positive and trusting environ-
ments where taking risk is not personally risky), may be evident only when
engagement is measured by the ISA Scale and, more specifically, by the social
dimension of that scale.
Although this study focused on differences in the measurement of
employee engagement, important relationships with workplace stress and
burnout were also revealed. It is clear that workplace stress and burnout nega-
tively affect employee engagement, as seen by both overall measures employed
in this study and by most of the subscales within each measure. Also, these
negative relationships are particularly strong with the emotional or affective
subscales of engagement. These dimensions are predicated on the presence of
enthusiasm, optimism, and positive emotions at work. So, in accordance with
COR theory, workplace interventions designed to reduce the negative effects
of stress on engagement might include frequent and constructive feedback,
team-building exercises, and the establishment of clear goals and celebrations
of success (Hobfoll, 2011). Additional interventions that lessen the likelihood
of emotional exhaustion (such as social support programs) may inhibit the
development of burnout and also improve the resilience of employee engage-
ment (Halbesleben, 2006; Kahn, 2007).
With respect to the population of employees contemplated by this
study, additional considerations may be warranted related to interventions

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


190 Anthony-McMann, Ellinger, Astakhova, Halbesleben

targeted at stress reduction (Hobfoll, 1989, 2011). In accordance with COR


theory, role conflict, in which individuals perceive conflicting demands at
work (such as IT professionals working to meet federally mandated dead-
lines in order to maximize federal incentive dollars while trying to ensure
the adequacy and thoroughness of their testing efforts) represents a clear
stress and might be mitigated by better resource allocation, improved
communications, and the development of a trusting culture in which IT
professionals feel safe to express their concerns. Role ambiguity, in which
individuals perceive that they lack the skills or information to adequately
do their jobs (such as IT professionals working to optimize a care delivery
process) represents yet another stress inducer and might be addressed by
teaming IT professionals with clinicians who better understand clinical pro-
cesses. Finally, role–stress fit, in which job roles become misaligned with
expected stresses (such as IT professionals facing the wrath of physicians
who are frustrated with new federal requirements), represents a significant
resource loss that is associated with feelings of emotional exhaustion and
depersonalization (LeRouge, Nelson, & Blanton, 2006; Schaufeli & Buunk,
2002). This, too, might be mitigated by the establishment of organizational
norms in which IT professionals are empowered and encouraged to request
assistance from management in cases where they fear confrontation or other
misalignment.
Limitations and Suggestions for Future Research
The use of self-report data introduces the possibility of bias due to common
method variance and, as a cross-sectional study, there is also some risk that
other explanations for the observed relationships are possible. Although the
findings from this study may be generalizable to the approximately 100,000
IT professionals working in U.S. community hospitals, the very specific con-
text of this study clearly limits its generalizability. Future studies should con-
firm the evidenced relationships with those working in other industries, other
jobs, different countries, and/or under different circumstances.
One of the major assumptions of this study is that the relationship
between workplace stress and employee engagement is linear. However,
there may be certain circumstances in which workplace stress and employee
engagement exhibit a curvilinear relationship, and this phenomenon may
represent an interesting topic for future research. Additional suggestions
for research include: (a) further analyzing the ISA and Rich scales at the
subscale level to provide a finer-grained analysis to identify any potential
methodological differences; (b) further examining the ISA Scale to better
understand the relationships between its social dimension and the other
ISA dimensions given the low correlations of this dimension with the other
two in this study—this instrument is relatively new, and it is possible that
the low correlation may be related to the context of this study, which differs
considerably from the manufacturing and retail contexts used in the initial

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


Exploring Different Operationalizations of Employee Engagement 191

and subsequent validation studies by Soane et al. (2012); (c) extending the
comparison of engagement measurement instruments to other published
instruments; (d) examining the specific influence of positive interper-
sonal relationships at work on employee engagement; and perhaps most
importantly, (e) conducting longitudinal studies that examine the impact
on employee engagement before and after specific, actionable, measur-
able workplace interventions that are targeted to improve levels of such
engagement.

Conclusion
The study of employee engagement has expanded significantly since Kahn’s
(1990) original conceptualization. As scholars continue to confirm both the
construct validity and the uniqueness of engagement in comparison to other,
better-known job-related constructs such as job satisfaction, job involve-
ment, and organizational commitment (Nimon, Shuck, & Zigarmi, 2014;
Shuck, Zigarmi, & Nimon, 2014), its utility as a measurable organizational
outcome will continue to increase. Indeed, HRD practitioners are particularly
interested in engagement-related concerns in health care—an industry that
has provided and continues to provide a rich environment for the study of
organizational change and the consequential impact on employee engage-
ment—but their interests also extend to other industries, work contexts,
and countries (Albrecht, 2010; Truss, Delbridge, Kerstin, Shantz, & Soane,
2014).
Some argue that the uncertainties and changes currently facing hospitals
and their employees—pressure to downsize, cultural impact of mergers, adop-
tion of disruptive technologies, reorganization of key processes—are similar
to those facing many, if not most, other industries, employers, and employees.
It is both intuitive and the contention of numerous researchers that engaging
the workforces of health care organizations and, by extension, all organiza-
tions, will be one of the key human resource–related strategic imperatives
necessary to successfully accomplish these change initiatives (Albrecht, 2010;
Kahn, 2010). To the extent that this study contributes to engagement litera-
ture by explaining some of the differences in how employee engagement is
measured and how those differences may impact organizational workforce
efforts, it helps all parties seeking to operationalize this phenomenon. We
are suggesting that disagreements about the differences in the operationaliza-
tion of employee engagement are less important than furthering collective
understanding about the differences so as to both clarify theory and inform
practitioners about the use of the best instrument(s) to match their organiza-
tional objectives. However, we maintain that ‘engagement’ is not a generalized
term, inclusive of all definitions and operationalizations, and its usefulness as
an actionable organizational phenomenon depends on clarity of meaning and
precision in measurement.

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


192 Anthony-McMann, Ellinger, Astakhova, Halbesleben

Acknowledgments
The authors would like to thank the editors, the associate editor assigned to
this manuscript, and anonymous reviewers for their insightful and helpful
comments on earlier versions of this article. The authors would also like to
acknowledge and thank the information technology professionals employed
by the community hospitals who participated in the original dissertation
study, entitled, “The Meaning and Measurement of Employee Engagement:
Exploring Different Operationalizations of Employee Engagement and Their
Relationships with Workplace Stress and Burnout Among IT Professionals
in Community Hospitals.” This dissertation, authored by Paula E. Anthony-
McMann, received the Esworthy Malcolm S. Knowles Outstanding Dissertation of
the Year Award in February, 2015.

References
Albrecht, S. L. (Ed.) (2010). The handbook of employee engagement: Perspectives, issues, research and
practice. Northhampton, MA: Edward Elgar.
Anthony-McMann, P. E. (2014). The meaning and measurement of employee engagement:
Exploring different operationalizations of employee engagement and their relationships with
workplace stress and burnout among IT professionals in community hospitals. (Doctoral dis-
sertation). The University of Texas at Tyler, Tyler, TX. Retrieved from http://scholarworks.
uttyler.edu/cgi/viewcontent.cgi?article=1002&context=hrd_grad
Armstrong, J. S., & Overton, T. S. (1977). Estimating nonresponse bias in mail surveys. Journal
of Marketing Research, 14(3), 396–402.
Bakker, A. B., Schaufeli, W. B., Leiter, M. P., & Taris, T. W. (2008). Work engagement: An
emerging concept in occupational health psychology. Work & Stress, 22( 3 ), 187 – 200 .
doi:10.1080/02678370802393649
Baron, R. M., & Kenny, D. A. (1996). The moderator-mediator variable distinction in social psy-
chological research: Conceptual, strategic, and statistical considerations. Journal of Personality
and Social Psychology, 31(6), 1173–1182.
Bertrand, M., & Mullainathan, S. (2001). Do people mean what they say? Implications for sub-
jective survey data. American Economic Review, 91(2), 67–72.
Blumenthal, D., & Tavenner, M. (2010). The “Meaningful Use” regulation for electronic health
records. New England Journal of Medicine, 363(6), 501–504.
Brooks, R., & Grotz, C. (2010). Implementation of electronic medical records: How health-
care providers are managing the challenges of going digital. Journal of Business & Economics
Research, 8(6), 73–84.
Brown, S. P., & Leigh, T. W. (1996). A new look at psychological climate and its relationship to
job involvement, effort and performance. Journal of Applied Psychology, 81(4), 358–368.
Christian, M. S., Garza, A. S., & Slaughter, J. E. (2011). Work engagement: A quantitative review
and test of its relations with task and contextual performance. Personnel Psychology, 64(1),
89–136. doi:10.1111/j.1744-6570.2010.01203.x
Cole, M. S., Walter, F., Bedeian, A. G., & O’Boyle, E. B. (2012). Job burnout and employee
engagement: A meta-analytic examination of construct proliferation. Journal of Management,
38(5), 1550–1581.
Demerouti, E., & Bakker, A. B. (2008). The Oldenburg Burnout Inventory: A good alternative
to measure burnout and engagement. In J. R. B. Halbesleben (Ed.), Handbook of stress and
burnout in health care. Hauppauge, NY: Nova Science.

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


Exploring Different Operationalizations of Employee Engagement 193

DesRoches, C. M., Campbell, E. G., Vogeli, C., Zheng, J., Rao, S. R., Shields, A. E., … Jha, A. K.
(2010). Electronic health records’ limited successes suggest more targeted uses. Health Affairs,
29(4), 639–646. doi:10.1377/hlthaff.2009.1086
Fletcher, L., & Robinson, D. (2013). Measuring and understanding engagement. In C. Truss, K.
Alfes, R. Delbridge, A. Shantz, & E. Soane (Eds.), Employee engagement in theory and practice.
New York, NY: Routledge.
Gan, T., & Gan, Y. (2013). Sequential development among dimensions of job burnout and
engagement among IT employees. Stress and Health, 30(2), 122–133.
Gorgievski, M. J., & Hobfoll, S. E. (2008). Work can burn us out or fire us up: Conservation of
resources in burnout and engagement. In J. R. B. Halbesleben (Ed.), Handbook of stress and
burnout in health care (pp. 7–22). New York, NY: Nova Science.
Hair, J. F., Black, W. C., Babin, B. J., & Anderson, R. E. (2010). Multivariate data analysis. Upper
Saddle River, NJ: Prentice Hall.
Halbesleben, J. R. B. (2006). Sources of social support and burnout: A meta-analytic test of the
conservation of resources model. Journal of Applied Psychology, 91(5), 1134–1145.
Halbesleben, J. R. B. (2010). A meta-analysis of work engagement: Relationships with burnout,
demands, resources, and consequences. In B. B. Arnold & M. P. Leiter (Eds.), Work engage-
ment: A handbook of essential theory and research (pp. 102–117). New York, NY: Psychology
Press.
Harter, J. K., Schmidt, F. L., Agrawal, S., & Plowman, S. K. (2013). The relationship between
engagement at work and organizational outcomes: 2012 Q12 meta analysis. Washington, DC:
Gallup.
Harter, J. K., Schmidt, F. L., & Hayes, T. L. (2002). Business-unit level relationship between
employee satisfaction, employee engagement, and business outcomes: A meta-analysis. Jour-
nal of Applied Psychology, 87(2), 268–279.
Hobfoll, S. E. (1989). Conservation of resources: A new attempt at conceptualizing stress. Ameri-
can Psychologist, 44(3), 513–524. doi:10.1037/0003-066x.44.3.513
Hobfoll, S. E. (2011). Conservation of resource caravans and engaged settings. Journal of Occupa-
tional & Organizational Psychology, 84(1), 116–122. doi:10.1111/j.2044-8325.2010.02016.x
Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis:
Conventional criteria versus new alternatives. Structural Equation Modeling, 6(1), 1–55.
Kahn, W. A. (1990). Psychological conditions of personal engagement and disengagement at
work. Academy of Management Journal, 33(4), 692–724. doi:10.2307/256287
Kahn, W. A. (1992). To be fully there: Psychological presence at work. Human Relations, 45(4),
321–349. doi:10.1177/001872679204500402
Kahn, W. A. (2007). Meaningful connections: Positive relationships and attachments at work. In
J. E. Dutton & B. R. Ragins (Eds.), Exploring positive relationships at work (pp. 189–206). New
York, NY: Psychology Press.
Kahn, W. A. (2010). The essence of engagement: Lessons from the field. In S. L. Albrecht (Ed.),
Handbook of employee engagement (pp. 20–30). Cheltenham, England: Edward Elgar.
Keenoy, T. (2014). Engagement: A mumuration of objects? In C. Truss, R. Delbridge, K. Alfes,
A. Shantz, & E. Soane (Eds.), Employee engagement in theory and practice. New York, NY:
Routledge.
LeRouge, C., Nelson, A., & Blanton, J. E. (2006). The impact of role stress fit and self-esteem on
the job attitudes of IT professionals. Information & Management, 43(8), 928–938.
Macey, W., & Schneider, B. (2008). The meaning of employee engagement. Industrial and Orga-
nizational Psychology, 1(1), 3–30.
Mahmood, M. H., Coons, S. J., Guy, M. C., & Pelletier, K. R. (2010). Development and testing of
the Workplace Stressors Assessment Questionnaire. Journal of Occupational and Environmental
Medicine, 52(12), 1192–1200. doi:10.1097/JOM.0b013e3181fb53dc
Malach-Pines, A. (2005). The Burnout Measure, Short Version. International Journal of Stress
Management, 12(1), 78–88.

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


194 Anthony-McMann, Ellinger, Astakhova, Halbesleben

Maslach, C., & Jackson, S. E. (1981). The measurement of experienced burnout. Journal of
Organizational Behavior, 2(2), 99–113.
Maslach, C., Schaufeli, W. B., & Leiter, M. P. (2001). Job burnout. Annual Review of Psychology,
52(1), 397–422.
May, D. R., Gilson, R. L., & Harter, L. M. (2004). The psychological conditions of meaningful-
ness, safety, and availability and the engagement of the human spirit at work. Journal of Occu-
pational Psychology, 77(1), 11–37.
McGraw, K. O., & Wong, S. P. (1996). Forming inferences about some intraclass correlation
coefficients. Psychological Methods, 1(1), 30–46.
Nimon, K., Shuck, B., & Zigarmi, D. (2014). Latent semantic analysis: Examining manifest valid-
ity of employee engagement measures. Paper presented at the 2014 AHRD Conference in the
Americas, Houston, TX.
Pines, A., & Aronson, E. (1988). Career burnout: Causes and cures. New York, NY: Free Press.
Podsakoff, P. M., MacKenzie, S. B., Jeong-Yeon, L., & Podsakoff, N. P. (2003). Common method
biases in behavioral research: A critical review of the literature and recommended remedies.
Journal of Applied Psychology, 88(5), 879–903.
Preacher, K. J. (2002). Calculation for the test of the difference between two indepedent correlation
coeffcients. Retrieved from http://quantpsy.org
Preacher, K. J., & Hayes, A. F. (2004). SPSS and SAS procedures for estimating indirect effects in
simple mediation models. Behavior Research Methods, Instruments, & Computers, 36(4), 717–
731. doi:10.3758/BF03206553
Rich , B. L. , Lepine , J. A. , & Crawford , E. R. ( 2010 ). Job engagement: Antecedents and
effects of job performance. Academy of Management Journal, 53(3), 617–635. doi:10.5465/
amj.2010.51468988
Robinson, D., Perryman, S., & Hayday, S. (2004). The drivers of employee engagement (Report).
London, England: Institute for Employment Studies.
Russell, J. A., & Barrett, L. F. (1999). Core affect, prototypical emotional episodes, and other
things called emotion: Dissecting the elephant. Journal of Personality and Social Psychology,
76(5), 805–819.
Saks, A. M. (2006). Antecedents and consequences of employee engagement. Journal of Manage-
rial Psychology, 21(7), 600–619.
Saks, A. M., & Gruman, J. A. (2014). What do we really know about employee engagement?
Human Resource Development Quarterly, 25(2), 155–182.
Schaufeli, W. B., Bakker, A. B., & Salanova, M. (2006). The measurement of work engage-
ment with a short questionnaire. Educational and Psychological Measurement, 66(4), 701–716.
doi:10.1177/0013164405282471
Schaufeli, W. B., & Buunk, B. P. (2002). Burnout: An overview of 25 years of research and theo-
rizing. In M. J. Schabracq, J. A. M. Winnubst, & C. L. Cooper (Eds.), Handbook of work and
health psychology (2nd ed., pp. 383–425). New York, NY: Wiley.
Schaufeli, W. B., Enzmann, D., & Girault, N. (1993). Measurement of burnout: A review. In W.
B. Schaufeli, C. Maslach, & T. Marek (Eds.), Professional burnout: Recent developments in theory
and research (pp. 199–215). Philadelphia, PA: Taylor & Francis.
Schaufeli, W. B., Salanova, M., González-Romá, V., & Bakker, A. B. (2002). The measurement
of engagement and burnout: A two sample confirmatory factor analytic approach. Journal of
Happiness Studies, 3(1), 71–92.
Schneider, B., Macey, W. H., Barbera, K. M., & Young, S. A. (2010). The role of employee trust
in understanding employee engagement. In S. L. Albrecht (Ed.), Handbook of employee engage-
ment: Perspectives, issues, research and practice. Northhampton, MA: Edward Elgar.
Shirom, A. (2005). Reflections on the study of burnout. Work & Stress, 19(3), 263–270.
doi:10.1080/0267837050037664
Shrout, P. E., & Fleiss, J. L. (1979). Intraclass correlations: Uses in assessing rater reliability.
Psychological Bulletin, 86(2), 420–428. doi:10.1037/0033-2909.86.2.420

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq


Exploring Different Operationalizations of Employee Engagement 195

Shuck, B. (2011). Four emerging perspectives of employee engagement: An integrative lit-


erature review. Human Resource Development Review , 10 ( 3 ), 304 – 328 . doi: 10.1177/
1534484311410840
Shuck, B., Ghosh, R., Zigarmi, D., & Nimon, K. (2012). The jingle jangle of employee engage-
ment: Further exploration of the emerging construct and implications for workplace learning
and performance. Human Resource Development Review, 12(1), 11–35.
Shuck, B., Reio, T., & Rocco, T. S. (2011). Employee engagement: An examination of antecedant
and outcome variables. Human Resource Development International, 14(4), 427–445.
Shuck, B., Zigarmi, D., & Nimon, K. (2014). Untangling the jangle: Examining the common and
unique variance of engagement. Paper presented at the 2014 AHRD Conference in the Americas,
Houston, TX.
Society for Human Resource Management (SHRM). (2015). Employee job satisfaction and
engagement: Optimizing organizational culture for success. Retrieved from https://www.shrm.
org/Research/SurveyFindings/Documents/2015-Job-Satisfaction-and-Engagement-Report.pdf
Soane, E., Truss, C., Alfes, K., Shantz, A., Rees, C., & Gatenby, M. (2012). Development and
application of a new measure of employee engagement: The ISA engagement scale. Human
Resource Development International, 15(5), 529–547.
Truss, C., Delbridge, R., Kerstin, A., Shantz, A., & Soane, E. (Eds.). (2014). Employee engagement
in theory and practice. New York, NY: Routledge.
Tsai, W., & Ghoshal, S. (1998). Social capital and value creation: The role of intrafirm networks.
Academy of Management Journal, 41(4), 464–476.
Viljevac, A., Cooper-Thomas, H. D., & Saks, A. M. (2012). An investigation into the validity
of two measures of work engagement. International Journal of Human Resource Management,
23(17), 3692–3709. doi:10.1080/09585192.2011.639542

Dr. Paula E. Anthony-McMann, PhD, Tyler, Texas, currently serves as the vice president
and CIO of ETMC Regional Healthcare System (ETMC). She is also the president and
CEO of HealthFirst, a wholly owned subsidiary of ETMC, and an adjunct professor in the
Department of Management and Marketing in the College of Business and Technology at
The University of Texas at Tyler.

Andrea D. Ellinger, PhD, is a professor in the Department of Human Resource


Development in the College of Business and Technology at The University of Texas at Tyler.

Marina Astakhova, PhD, is an associate professor in the Department of Management and


Marketing in the College of Business and Technology at The University of Texas at Tyler.

Jonathon R. B. Halbesleben, PhD, is the senior associate dean and Russell Professor of
Business Administration for the Culverhouse College of Commerce at the University of
Alabama.

Corresponding Author:
Paula Anthony-McMann can be contacted at panthony117@gmail.com

HUMAN RESOURCE DEVELOPMENT QUARTERLY • DOI: 10.1002/hrdq