CC BY-NC-ND
0tli^^^M AI IIIII.E IN PRESS
Burnout Researchxxx (2015) xxx-xxx
ELSEVIER
Contents lists available at ScienceDirect
Burnout Research
journal homepage www.elsevier.com/locate/burn
Revisiting the interplay between burnout and work engagement: An Exploratory Structural Equation Modeling (ESEM) approach^
Q3 Sarah-Geneviève Trépanier2'*, Claude Fernetb, Stéphanie Austinb, Julie Ménarda
a Department of Psychology, Université du Québec à Montréal, Quebec, Canada b Department of Management, Université du Québec à Trois-Rivières, Quebec, Canada
ARTICLE INFO
Article history: Received 28 October 2014 Received in revised form 13 April 2015 Accepted 14 April 2015
Keywords: Burnout
Work engagement
Job Demands-Resources QD-R) model Exploratory Structural Equation Modeling (ESEM)
ABSTRACT
This study aimed to investigate the interplay between burnout and work engagement. More specifically, we examined the energy and identification continua theorized to underlie the relationship between burnout and work engagement by simultaneously evaluating the factorial structure of the Maslach Burnout Inventory-General Survey (MBI-GS) and the Utrecht Work Engagement Scale (UWES). Results from Exploratory Structural Equation Modeling (ESEM) offered little support for these continua, suggesting that burnout and work engagement are not diametrical counterparts. Moreover, ESEM significantly altered the relationships burnout and work engagement hold with job demands and resources (i.e., work overload, job autonomy, and recognition), as well as health-related (i.e., psychological distress) and motivational (i.e., turnover intention) outcomes. These findings shed new light on the health-impairment and motivational processes theorized by the JD-R model.
© 2 015 Published by Elsevier GmbH. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
1. Introduction
The field of positive psychology has greatly influenced our current conceptualization of employee functioning by highlighting the importance of not only preventing negative manifestations (i.e., ill-being) but also promoting positive ones (i.e., well-being). Due to this conceptual shift, occupational health researchers and practitioners investigating burnout-a key indicator of employee ill-being-have expanded their scope of interest and begun focusing on burnout's antipodal counterpart, work engagement. It has been suggested that the dimensions of burnout and work engagement represent opposite ends of two continua reflecting employees' overall level of energy and identification with their work (Bakker, Schaufeli, Leiter, &Taris, 2008; Demerouti&Bakker, 2008). Because this proposition has not been subjected to an extensive empirical testing, we attempted to investigate this issue. More specifically, we investigated the energy and identification continua proposed to underlie the relationship between burnout and work engagement by simultaneously evaluating the factorial structure of the Maslach
* This work was supported by a grant from the Fonds Québécois de la Recherche sur la Société et la Culture awarded to Claude Fernet.
* Corresponding author at: Department of Psychology, Université du Québec à Montréal, C.P. 8888, Succ. Centre-Ville, Montréal, Québec, Canada H3C 3P8.
Tel.: +1 514 987 3000x5331.
E-mail address: trepanier.sarah-genevieve@uqam.ca (S.-G. Trépanier).
Burnout Inventory-General Survey (MBI-GS) and the Utrecht Work Engagement Scale (UWES) using a novel statistical approach called exploratory structural equation modeling (ESEM). We also examined the health-impairment and motivational processes proposed by the Job Demands-Resources (JD-R) model (Demerouti, Bakker, Nachreiner, & Schaufeli, 2001). Through SEM with ESEM factors of burnout/work engagement, we evaluate the pattern of relationships between job characteristics (job demands and resources), the dimensions of burnout/work engagement, as well as health-related and motivational outcomes.
1.1. Burnout and work engagement: conceptualization and measurement
Extensive research conducted on burnout over the course of more than 30 years has improved our understanding of its nature. Burnout can be viewed as a negative psychological response resulting from employees' interaction with their job (Leiter & Bakker, 2010; Maslach, 1982). This negative reaction is said to manifest itself through two core dimensions: emotional exhaustion and cynicism (Bakker, Demerouti, & Sanz-Vergel, 2014). Emotional exhaustion reflects feelings of being overextended and drained of one's mental, emotional and physical resources, whereas cynicism is characterized by an overly negative and detached attitude regarding one's work. Of the instruments developed to measure burnout the Maslach Burnout Inventory-General Survey (MBI-GS;
60 61 62
http://dx.doi.org/10.1016/j.burn.2015.04.002
2213-0586/© 2015 Published by Elsevier GmbH. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
'lIM^^^M AHIHI.E IN PRESS
S.-G. Trépanier et al. / Burnout Research xxx (2015) xxx-xxx
63 Schaufeli, Leiter, Maslach, & Jackson, 1996) is the most widely used
64 scale (Schaufeli & Taris, 2005).
65 More recently, researchers have begun to investigate employee
66 psychological functioning from a more positive perspective:
67 work engagement. Work engagement can be defined as "a pos-
68 itive, fulfilling, work-related state of mind" (Schaufeli, Salanova,
69 Gonzalez-Roma, & Bakker, 2002, p. 74). More specifically, when
70 experiencing work engagement, employees exhibit high levels of
71 vitality, and willingness to fully invest themselves in their tasks (i.e.,
72 vigor). They also have a strong sense of involvement and enthusi-
73 asm regarding their work (i.e., dedication). Work engagement is
74 most commonly measured using the Utrecht Work Engagement
75 Scale (UWES; Schaufeli et al., 2002).
76 1.1.1. The relationship between burnout and work engagement
77 Theoretically, vigor and dedication are considered to be the
78 direct opposites of emotional exhaustion and cynicism, respec-
79 tively (Schaufeli et al., 2002). As such, emotional exhaustion and
80 vigor are viewed as opposite ends of an underlying continuum
81 labeled "energy", whereas cynicism and dedication are viewed as
82 opposite ends of an underlying continuum labeled "identification".
83 In this perspective, burnout and work engagement are considered
84 as opposite sides of the same coin and not independent constructs
85 (Gonzalez-Roma, Schaufeli, Bakker, & Lloret, 2006). This implies
86 that employees who score high on one dimension of a contin-
87 uum (e.g., dedication) would necessarily score low on the other
88 end of that continuum (e.g., cynicism). However, very few studies
89 (e.g., Demerouti, Mostert, & Bakker, 2010; Gonzalez-Roma et al.,
90 2006; Makikangas et al., 2012) have adequately investigated this 9iQ5 proposition empirically. For example, Demerouti et al. (2010) con-
92 ducted confirmatory factor analysis (CFA) using the MBI-GS, the
93 UWES and the Oldenburg Burnout Inventory (OLBI), which con-
94 tains positively and negatively worded items said to reflect both
95 ends of the energy (i.e., labeled exhaustion and vigor) and identifi-
96 cation (i.e., labeled disengagement and dedication) continua. They
97 found that the identification dimensions (cynicism/disengagement
98 and dedication) represent identical second order factors, suggest-
99 ing that they can be considered as opposite ends of a single
100 continuum. In this view, negative and excessively detached atti-
101 tudes about one's work (i.e., cynicism) and a strong involvement
102 in one's work (i.e., dedication) would reflect diametrically oppo-
103 site attitudes. However, the energy dimensions (exhaustion and
104 vigor) were found to represent independent second order fac-
105 tors: exhaustion (i.e., feelings of being overextended) and vigor
106 (i.e., high levels of energy and mental resilience while working)
107 appear to be distinct, albeit highly related (r =.87), experiences.
108 More recently, Makikangas et al. (2012) investigated intraindivid-
109 ual developmental patterns of burnout and work engagement (and
110 their interplay) in a two-year follow-up study among managers.
111 Results showed that managers who belonged to the category "low
112 cynicism" also predominantly belonged to the "stable high dedica-
113 tion" category, supporting the identification continuum. Much like
114 Demerouti et al. (2010), Makikangas et al. (2012) found little sup-
115 port forthe energy continuum: managers'experiences of emotional
116 exhaustion and vigor appeared to evolve independently.
117 The fact that past research has failed to provide unambiguous
118 empirical support for the two continua assumed to underlie the
119 relationship between burnout and work engagement can be partly
120 explained from a statistical standpoint. Like Demerouti et al. (2010)
121 most studies (e.g., Hakanen, Bakker, & Schaufeli, 2006; Hakanen,
122 Schaufeli, & Ahola, 2008; Schaufeli & Bakker, 2004; Schaufeli et al.,
123 2002) have investigated the relationships between burnout and
124 work engagement (and their dimensions) using CFA. However,
125 given the rigidity of some of its fundamental postulates (e.g.,
126 strict requirement of zero cross-loadings), CFA measurement mod-
127 els may not be the most suitable approach for investigating the
relationship between concepts that are theoretically very closely 128
related, such as burnout and work engagement. A relatively new 129
statistical tool called Exploratory Structural Equation Modeling 130
(ESEM; Asparouhov & Muthen, 2009) may provide the flexibility 131
needed to conduct a more thorough investigation of the interplay 132
between the dimensions of burnout and work engagement and 133
their potential underlying continua. 134
1.1.2. Investigating burnout and work engagement: ESEM versus 135 CFA 136
CFA is a statistical approach often used in occupational health 137
psychology to assess latent constructs (e.g., job demands, moti- 138
vation, and work engagement). In CFA measurement models, 139
researchers specify (1) the number of factors assumed to reflect 140
the latent constructs and (2) which items (or indicators) repre- 141
sent each factor. Items are specified to represent their factor only: 142
all cross-links are fixed at zero. However, the no cross-loading 143
assumption is often too restrictive and may provide a biased rep- 144
resentation of the relationship between theoretically related latent 145
factors (e.g., dedication and cynicism) by overestimating the corre- 146
lation between these factors (Asparouhov & Muthen, 2009; Marsh 147
et al., 2009; Morin, Marsh, & Nagengast, 2013). These overesti- 148
mated correlations may result in a distorted representation of the 149
structural relationship between the latent factors and other con- 150
structs (e.g., work-related antecedents and outcomes of burnout 151
and work engagement) when integrated in structural equation 152
modeling (SEM; Asparouhov & Muthen, 2009). 153
ESEM may allow scholars to overcome the limits associated with 154
CFA. This modeling procedure enables researchers to freely esti- 155
mate all cross-loadings of indicators of latent factors Much recent 156
research has illustrated the merits of ESEM over CFA (e.g., Guay, 157
Morin, Litalien, Valois, & Vallerand, 2015; Marsh, Liem, Martin, 158
Morin, & Nagengast, 2011; Marsh, Nagengast, & Morin, 2012). The 159
common denominator of these studies is that they reveal that 160
allowing cross-loading between theoretically linked factors (via 161
ESEM) provides a significantly better representation of the data 162
than constraining all cross-loadings at zero (via CFA). Moreover, 163
the inter-correlations between latent factors as well as the corre- 164
lations between these factors and other variables (i.e., theoretical 165
antecedents or outcomes) are considerably reduced in ESEM solu- 166
tions. For example, in a multi-sample study conducted among 167
students, Guay et al. (2015) found the inter-relationships between 168
types of motivation (e.g., intrinsic, extrinsic) to be considerably 169
lower in the ESEM solution (r =.24-.46) than in the CFA solu- 170
tion (r = .56-.80). Moreover, these types of motivation were more 171
strongly related to perceived academic competence (i.e., a theo- 172
retical antecedent) in the CFA measurement models of motivation 173
(r = -.58-.57), compared to the ESEM solutions (r = -.35-.32). Over- 174
all, these findings highlight that, due to its restrictive nature, CFA 175
measurement models may result in a biased representation of the 176
relationship between strongly theoretically related concepts by 177
artificially inflating these relationships. 178
1.1.3. Burnout and work engagement: associations with job 179 characteristics and outcomes 180
The JD-R model (Demerouti et al., 2001; Schaufeli & Bakker, 181
2004) describes the psychological processes through which job 182
characteristics (i.e., demands and resources) act as key predictors 183
of burnout and work engagement. Accordingly, in the health- 184
impairment process, job demands (i.e., negatively valued aspects 185
of the job that require sustained effort; Schaufeli & Taris, 2014) 186
deplete employees' mental, emotional and physical resources and 187
therefore lead to burnout (Bakker & Demerouti, 2007; Schaufeli 188 & Taris, 2014). The prolonged experience of burnout results in Q6 189
negative health consequences (Bakker & Demerouti, 2007; Bakker 190
et al., 2014), including psychosomatic complaints and depressive 191
tli^^^M AI IIHI.E IN PRESS
S.-G. Trépanier et al. / Burnout Research xxx (2015) xxx-xxx
192 symptoms (Hakanen et al., 2008,2006; Korunka, Kubicek, Schaufeli,
193 & Hoonakker, 2009; Schaufeli & Bakker, 2004). Conversely, the
194 motivational process is proposed to underlie the relationship
195 between job resources, work engagement and indicators of psycho-
196 logical investment at work (Bakker et al., 2014). Job resources are
197 positively valued aspects of the job that help employees achieve
198 work goals, alleviate the strain associated with job demands and
199 stimulate personal development and growth (Bakker & Demerouti,
200 2007; Schaufeli and Taris, 2014). Given these positive effects, job
201 resources (e.g., social support, performance feedback) boost work
202 engagement (Bakker et al., 2014; Halbesleben, 2010; Schaufeli &
203 Bakker, 2004) and lead to various positive motivational outcomes
204 such as organizational commitment and low turnover intention 2oQ7 (Hakanen etal., 2006; Korunaetal., 2008; Schaufeli & Bakker, 2004).
206 The health-impairment and motivational processes are pro-
207 posed to be relatively independent (Bakker & Demerouti, 2007),
208 although cumulative evidence suggests that they are interrelated
209 (Schaufeli and Taris, 2014). Indeed, job demands have been found to
210 be negatively related to work engagement, whereas job resources
211 have been negatively linked to burnout (e.g., Crawford, LePine, &
212 Rich, 2010; Hakanen et al., 2006). Furthermore, burnout has been
213 found to be negatively related to motivational outcomes, whereas
214 work engagement has been positively linked to health-related
215 outcomes (e.g., Bakker, Demerouti, & Schaufeli, 2003; Hakanen
216 et al., 2006; Richardsen, Burke, & Martinussen, 2006; Schaufeli &
217 Bakker, 2004). Nevertheless, these cross-links do not appear to be as
218 substantial or systematic as the direct links underlying the health-
219 impairment and motivational processes (Halbesleben, 2010; Hu,
220 Schaufeli, & Taris, 2011; Schaufeli & Bakker, 2004). For example, in
221 a two-sample study, Hu et al. (2011) found that the links between
222 job demands (e.g., workload, emotional demands) and burnout (.58
223 and .62) were stronger than those between job demands and work
224 engagement (-.09 and -.05). Conversely, the links between job
225 resources (i.e., job control, colleague support) and work engage-
226 ment (.47 and .53) were stronger than those between job resources
227 and burnout (-.18 and -.37).
228 However, because most studies investigating both processes
229 have used SEM with CFA factors of burnout and work engagement,
230 it is possible that their representation of the relationships between
231 job characteristics, burnout/work engagement, and health-related
232 as well as motivational outcomes is significantly biased (Marsh
233 et al., 2009). Indeed, given that CFA measurement models result
234 in substantially inflated factor correlations (e.g., between burnout
235 and work engagement), it is likely to distort subsequent structural
236 analyses. As such, because it provides a more exact estimate of
237 the relationship between burnout and work engagement, ESEM
238 should also provide a more accurate representation of the relation-
239 ship between these two concepts and job demands, job resources,
240 as well as health-related and motivational indicators of employee
241 functioning. Investigating the health-impairment and motivational
242 processes through ESEM would thus shed new light on the possible
243 interdependency of the two processes.
244 1.2. The present study
245 The aim of this study is to deepen our understanding of the inter-
246 play between burnout and work engagement by delving further
247 into the energy and identification continua. First, we investigate
248 the factorial structure of the MBI-GS and the UWES simultaneously
249 using ESEM. This approach allows indicators to cross-load on mul-
250 tiple factors. As such, items representing one end of a continuum
251 should also load on the factor representing the opposite end of
252 that continuum (albeit negatively). More specifically, based on the
253 above mentioned JD-R-based empirical evidence and theoretical
254 propositions in support of the energy and identification continua,
255 we propose the following hypotheses:
Hypothesis 1. Emotional exhaustion items will have significant 256
cross-loadings on the vigor factor and vigor items will have signifi- 257
cant cross-loadings on the emotional exhaustion factor (supporting 258
the energy continuum). 259
Hypothesis 2. Cynicism items will have significant cross-loadings 260
on the dedication factor and dedication items will have significant 261
cross-loadings on the cynicism factor (supporting the identification 262
continuum). 263
Hypothesis3. Atwo-factorESEMsolution(i.e.,emotionalexhaus- 264
tion/vigor and cynicism/dedication) representing the two continua 265
will provide a better fit to the data than a four-factor ESEM solu- 266
tion (i.e., emotional exhaustion, cynicism, vigor and dedication) 267
representing the four separate dimensions. 268
By exploring whether burnout and work engagement are dia- 269
metrical counterparts, this study will evaluate the added value 270
of (or redundancy in) investigating both work engagement and 271
burnout to assess employees' level of energy and identification with 272
their work. 273
Second, we conduct exploratory ESEM analyses to examine the 274
health-impairment and motivational processes proposed by the 275
JD-R model. By comparing two structural models (one with CFA 276
factors of burnout/work engagement and one with ESEM factors 277
of these constructs), this study ultimately aims to assess whether 278
ESEM significantly alters the pattern of relationship between job 279
characteristics, burnout/work engagement, and health-related and 280
motivational outcomes. 281
2. Method 282
2.1. Participants and procedures 283
The sample comprised school teachers (n = 1159, participation 284
rate of 39%) working in the province of Quebec, Canada. All teach- 285
ers received a letter at work describing the purpose of the study in 286
detail and inviting them to complete an online questionnaire. The 287
majority of participants were women (85.8%). Mean age was 27.79 288
years (SD = 4.13), with an average of3.29 (SD = 1.68) years of expe- 289
rience on the job. The majority taught in primary schools (60.3%), 290
34.7% in secondary schools and 5% in other school settings. 291
2.2. Measures 292
All measures were administered in French. Means, standard 293
deviations and latent correlations of these measures are presented 294
in Table 1. Reliability of the measures was established by Hancock's 295
coefficient (i.e., coefficient H; Hancock & Mueller, 2001), which uses 296
standardized factor loadings obtained through CFA measurement 297
models to estimate the stability of latent constructs across mul- 298
tiple observed variables. Values equal to or greater than .70 are 299
considered satisfactory (Hancock & Mueller, 2001). 300
2.2.1. Burnout 301 The core dimensions of burnout were assessed using the emo- 302
tional exhaustion and cynicism subscales of the MBI-GS (Schaufeli 303
et al., 1996). Each of these subscales contains five statements per- 304
taining to either emotional exhaustion (e.g., "I feel emotionally 305
drained by my work", coefficient H = .92) or cynicism (e.g., "I doubt 306
the significance of my work", coefficient H = .87). Participants were 307
asked to indicate how often they experienced these feelings at work 308
on a scale from 1 (never) to 7 (every day). 309
2.2.2. Work engagement 310 The core dimensions of work engagement were assessed using 311
the vigor and dedication subscales of the UWES (Schaufeli et al., 312
fli^^^M AllIil.E IN PRESS
4 S.-G. Trepanier et al. / Burnout Research xxx (2015) xxx-xxx
Table 1
Means, standard deviations and correlations between latent variables.
Mean SD Range Emotional Cynicism Vigor Dedication Work Job Recognition Turnover
exhaustion overload autonomy intention
Emotional 3.168 1.287 1-7 -
exhaustion
Cynicism 2.331 1.097 1-7 .467 .740 -
Vigor 5.388 1.053 1-7 -.265 -.565 -.356 -.706 -
Dedication 5.675 1.001 1-7 -.170 -.530 -.436 -.764 .580 .924 -
Work overload 3.078 .808 1-5 .753 .743 .383 .494 -.251 -.381 -.263 -.369 -
Job autonomy 3.452 .677 1-5 -.436 -.506 -.422 -.538 .443 .550 .447 .516 -.603 -.604 -
Recognition 3.733 .832 1-5 -.255 -.480 .384 .456 -.321 .544 -
-.361 -.534 .492 .520 -.322 .544
Turnover 1.871 1.372 1-7 .488 .757 -.432 -.514 .422 -.420 -.447
intention .610 .793 -.599 -.662 .426 -.420 -.447
Psychological 1.688 .527 1-4 .667 .531 -.390 -.337 .538 -.449 -.372
distress .713 .622 -.503 -.478 .538 -.449 -.372
Note. Correlations of the ESEM solution are in bold. Means and SD obtained through CFA measurement models.
2002). Sample items are: "At work I feel like I am bursting with energy" (vigor; 6 items; coefficient H =.89) and "I am enthusiastic about my work" (dedication; 5 items; coefficient H =.93). Participants were asked to indicate how often they experienced these feelings at work on a scale from 1 (never) to 7 (every day).
2.2.3. Job characteristics
Job demands were assessed with the work overload subscale of the Areas of Work Life Scale (AWS; Leiter & Maslach, 2004), whereas job resources were assessed with the job autonomy and recognition subscales of the AWS. Sample items are: "I do not have time to do the work that must be done" (work overload; 6 items; coefficient H = .88), "I have control over how I do my work" (job autonomy; 3 items; coefficient H = .58) and "My work is appreciated" (recognition; 4 items; coefficient H = .89). Participants were asked to rate the frequency with which they experienced these situations on a five-point scale from 1 (never) to 5 (almost always). In the SEM analyses, the items of the three subscales were used as indicators of their respective latent factor.
2.2.4. Psychological distress
The French version (Préville, Boyer, Potvin, Perrault, & Légaré, 1992) of the Psychiatric Symptom Index (PSI; Ilfeld, 1976) was used to assess psychological distress, a health-related outcome of burnout and work engagement. This scale measures the presence of anxiety (3 items; coefficient H =.83) and depressive (5 items; coefficient H = .82) symptoms, as well as irritability (4 items; coefficient H= .88) and cognitive problems (2 items; coefficient H= .88) experienced during the previous week. A sample item of the scale is: "I felt easily annoyed or irritated" (i.e., irritability problem). Items were scored on a four-point scale ranging from 1 (never) to 4 (very often). In the SEM analyses, mean scores on the four subscales were used as indicators of the latent construct of psychological distress.
2.3. Ethical considerations
Approval for the study was obtained from the research ethics board of the researchers' institution. All participants received a letter explaining the purpose (i.e., investigating workplace factors associated with well-being in the teaching profession) and a description of what their participation consisted of (i.e., taking about 30 min to complete an online questionnaire regarding their work experiences). The confidentiality and anonymity of responses were also emphasized in the letter. No incentive was given in exchange for participation.
2.4. Statistical analyses
In the present study, all analyses were performed using Mplus (Muthen & Muthen, 2012) with the WLSMV estimator for categorical variables. CFA and ESEM measurement models were tested to investigate the factorial structure of the MBI-GS and UWES (measurement analyses). For each analysis (CFA and ESEM), two types of measurement models were tested: a two-factor and a four-factor structure. In the CFA solution, each indicator of the MBI-GS and UWES was allowed to load on its respective factor only. Latent factors were allowed to correlate. In the two ESEM solutions (a two- and a four-factor structure) all loadings were freely estimated using an oblique Geomin rotation (the default rotation solution in Mplus) with an epsilon value of 0.5 (Marsh et al., 2009, 2012). The latent factors were also allowed to correlate. The goodness-of-fit of all tested models was evaluated using three fit indices compatible with the WLSMV estimator: the Comparative Fit Index (CFI), the Tucker-Lewis Index (TLI) and the Root Mean Square Error of Approximation (RMSEA). Values higher than .95 for the CFI and TLI indicate a good fit (Hooper, Coughlan, & Mullen, 2008; Hu & Bentler, 1999). For the RMSEA, values lower than .07 (with the upper limit of the confidence interval [CI] less than .08) represent reasonable error of approximation (Hooper et al., 2008; Steiger, 2007).
2.2.5. Turnover intention
Turnover intention was evaluated as a motivational outcome of burnout and work engagement, using three items adapted from O'Driscoll and Beehr's scale (1994; e.g., "I plan on looking for another job within the next 12 months"). Items were scored on a scale ranging from 1 (strongly disagree) to 7 (strongly agree). In the SEM analyses, each item was used as an indicator of the latent construct of turnover intention (coefficient H= .97).
3. Results
3.1. Measurement analyses
Two CFA measurement models of the MBI-GS and UWES were tested: (1) a four-factor (M1) structure (i.e., emotional exhaustion, cynicism, vigor, and dedication) and (2) a two-factor (M2) structure comprised of one factor including the emotional exhaustion and
tli^^^M AI IIHI.E IN PRESS
S.-G. Trépanier et al. / Burnout Research xxx (2015) xxx-xxx
Table 2
Fit indices for the tested models.
Model description
RMSEA and 90% CI
Ml vs M2
M3 vs Ml M3 vs M4
M5 vs M6
4672.316*'
994.896*' 1627.100*'
1102.506*'
CFA measurement models
M1: Four factors 1934.131 183 .962 .957 .099 (.095-.103)
M2: Two factors (energy and identification continua) 6606.447 188 .862 .846 .189 (.183-.191)
ESEM measurement models
M3: Four factors 939.235 132 .983 .972 .079 (.075-.084)
M4: Two factors 2566.335 169 .948 .936 .121 (.117-.125)
Structural Analyses
M5:SEM with ESEM factors of burnout/work engagement 3066.010 698 .963 .956 .057 (.055-.060)
M6: CFA with ESEM factors of burnout/work engagement 4168.516 749 .946 .941 .067 (.065-.069)
Note. CFI = comparative fit index; TLI = Tuckey-Lewis index; RMSEA=root mean square error of approximation; CI = confidence interval; SRMR=standardized root mean square; MC: model comparison; Ax2 = chi-square difference; ** p<.001.
Table 3
Measurement model; correlations between the dimensions of burnout and work engagement (CFA and ESEM solutions).
Emotional exhaustion Cynicism Vigor Dedication
Emotional exhaustion -
Cynicism .473/.740
Vigor -.280/-.561 -.352/-.704
Dedication -.188/-.530 -.473/-.764 .597/.924
Note. ESEM correlations are in bold above the dashed line.
391 vigor items (i.e., energy continuum) and a second factor including
392 the cynicism and dedication items (i.e., identification continuum).
393 Ml provided an adequate fit to the data with the exception of the
394 RMSEA which was above the .07 (and upper CI limit above the .08)
395 threshold (see Table 2). Ml also provided a significantly better fit
396 to the data than M2, which provided a poor fit to the data (Table 2).
397 Correlations between the four latent factors in Ml were high, rang-
398 ing from -.530 (between emotional exhaustion and dedication) to
399 .924 (between vigor and dedication; see Table 3).
400 Next, two ESEM measurement models with four (M3) and two
401 (M4) factors were tested (see Table 2). Results show that M4 did
402 not fit the data particularly well (none of the fit indices respected
403 their cut-off thresholds). M3 provided a satisfactory fit to the data.
404 Although the RMSEA was above .07, the upper CI limit was very
405 close to .08, suggesting a reasonable error of approximation. Results
406 also show that M3 provided a significantly better fit to the data than
407 M4. Moreover, M3 (i.e., ESEM four-factor solution) provided a sig-
408 nificantly better fit than Ml (i.e., CFA four-factor solution). Overall,
409 these results infirm Hypothesis 3. Correlations between the four
410 latent factors of M3 (ESEM four-factor solution) decreased signif-
411 icantly compared to Ml (CFA four-factor solution), ranging from
412 -.l88 to .597 (see Table 3). The strongest correlation was found
413 between vigor and dedication (r =.597), followed by the correla-
414 tion between emotional exhaustion and cynicism (r =.473). The
415 lowest correlations were between vigor and emotional exhaus-
416 tion (r= -.280) and between emotional exhaustion and dedication
417 (r= -.l88). Taken together, these results provide weak preliminary
418 support for the energy and identification continua. Indeed, results
419 showed that the four dimensions of burnout and work engagement
420 are best represented as distinct factors as opposed to components of
421 two underlying factors. Moreover, the pattern of correlations shows
422 that the components of the energy and identification continua are
423 not strongly negatively related. There is a small correlation between
424 emotional exhaustion and vigor (r = -.280; Cohen, l988) and a
425 moderate correlation between cynicism and dedication (r = -.473;
426 Cohen, l988).
427 All factor loadings (primary and cross-loadings) of the ESEM
428 four-factor solution are presented in Table 4. These factor loadings
429 enable a more rigorous evaluation of the energy and identification
430 continua. In order to support these continua, strong cross-loadings
431 for each latent factor should be found from indicators representing
the opposite dimension of the same continuum (e.g., dedication 432
items should have strong cross-links on the cynicism factor). The 433
results shown in Table 3 offer little support for any of the continua 434
(infirming Hypotheses l and 2). For the two dimensions of burnout, 435
no cross-loadings reached the threshold of .30 (Tabachnick & Fidell, 436
2007). This suggests that both latent factors are relatively distinct. 437
A different pattern of results was obtained for the two dimensions 438
of work engagement. Three out of five dedication items had cross- 439
loadings of .30 or higher on the vigor latent factor, and two out of 440
six vigor items had cross-loadings of .30 or higher on the dedication 441
latent factor.l 442
Overall, these results offer little empirical support for the energy 443
and identification continua and reveal that the two dimensions of 444
work engagement are highly intertwined. Moreover, these results 445
suggest that ESEM, which considers cross-loadings between the 446
core dimensions of burnout and work engagement, more ade- 447
quately reflects the interplay between these dimensions than CFA. 448
3.2. Structural analyses 449
Exploratory SEM analyses were conducted subsequently to 450
test the structural relationships between job characteristic (i.e., 451
work overload, job autonomy, and recognition), burnout/work 452
engagement, and health-related as well as motivational outcomes 453
(i.e., psychological distress and turnover intention). More specifi- 454
cally, two models were compared: M5 including ESEM-factors of 455
burnout/work engagement, and M6, including CFA-factors of these 456
constructs. In both models, all links between job characteristics 457
and the four dimensions of burnout/work engagement as well as 458
between these dimensions and the two outcomes were assessed. 459
l Given that the short version of the UWES (UWES-9; Schaufeli, Bakker, & Salanova, 2006) is often used to assess work engagement, CFA and ESEM measure-
ment analyses were subsequently conducted using the MBI-GS and the UWES-9. The results revealed a pattern similar to the one obtained with the complete version of the work engagement scale. More specifically, no significant (i.e., above .30) cross-loadings were observed for the latent factors of emotional exhaustion and cynicism. Two dedication items (out of three) had cross-loadings of .30 or higher on the vigor latent factor and one vigor item (out of three) had cross-loadings of .30 or higher on the dedication latent factor. Detailed results of the ESEM measurement model containing the UWES-9 can be obtained from the first author.
IM^^^M AlllHl.E IN PRESS
S.-G. Trépanier et al. / Burnout Research xxx (2015) xxx-xxx
Table 4
Measurement Model: CFA and ESEM solutions for the MBI-GS and UWES.
Factor loadings
Factor 1
Factor2
Factor3
Factor 4
Emotional exhaustion Item 1 .752/.843
Item 2 .932/.855
Item 3 .728/.852
Item 4 .615/.885
Item 5 .731/.88Î
Cynicism Item 1 .153
Item 2 .150
Item 3 .034
Item 4 .097
Item 5 .113
Vigor Item 1 -.146
Item 2 -.114
Item 3 .099
Item 4 .143
Item 5 -.058
Item 6 -.086
Dedication Item 1 -.060
Item 2 -.014
Item 3 -.080
Item 4 -.080
Item 5 -.029
-.029 .080 .248 .214
.702/.868 .704/.9Î9 .497/.497 .465/.637 .582/.745
-.205 -.137 .155 .098 .204 -.116
-.274 -.255 -.188 -.041 -.092
.013 .032 -.256 -.240 .002
-.153 .032 -.256 .006 .071
.420/.924 .738/.879 .402/.5Î7 .358/.36Î .357/.3Î3 .754/.898
.395 .428 .405 .129 .062
-.083 -.084 .064 .057 -.082
-.039 -.056 -.063 -.211 -.270
.369 .084 .392 .250 .147 .128
472/.948 471/.875 448/.893 733/.836 .857/.8Î9
Note. CFA loadings are in italics below the dashed line. Factor loadings of the items reflecting the theoretical counterpart of each factor are in bold. Significant cross-loadings are underlined.
460 Although M6 fits the data reasonably well (CFI and TLI were close
461 to the .95 threshold) the results indicate that M5 provided a sig-
462 nificantly better fit to the data (see Table 3). The results of both
463 solutions (ESEM and CFA) are depicted in Fig. 1. Results show that
464 the structural links between job characteristics and the dimensions
465 of burnout and work engagement in the CFA solution are generally
stronger (7 out of 10) than those obtained in the ESEM solution. 466
Interestingly, the CFA solution reveals a significant link between 467
job autonomy and cynicism that was not significant in the ESEM 468
solution. Moreover, the "emotional exhaustion-turnover intention" 469
and "dedication-psychological distress" relationships were found 470
to be non-significant in the CFA solution but significant in the ESEM 471
/ Emotional ^-"""TV exhaustion
.743/ ^ /
Fig. 1. Results comparing the structural relationships between job characteristics, the dimensions of burnout/work engagement and outcomes through CFA and ESEM. Note. Results of the ESEM solution are in bold above the dashed line.
fli^^^M AI IIHI.E IN PRESS
S.-G. Trépanier et al. / Burnout Research xxx (2015) xxx-xxx
472 solution. Subsequently, regression coefficient comparison (using
473 unstandardized coefficient estimates and standard errors) was con-
474 duct to more rigorously compare M5 (ESEM solution) and M6
475 (CFA solution). Results reveal several significant differences. These
476 differences are presented in bold dotted lines in Fig. 1. More specif-
477 ically, M6 revealed significantly stronger relationships between (1)
478 work overload and cynicism (z = 2 . 2 7), (2) work overload and ded-
479 ication (z = 2 .06), (3) job autonomy and dedication (z =2.0 2), (4)
480 recognition and vigor (z = 2 .54) and (5) recognition and dedication
481 (z = 4.37). On the other hand, M5 revealed significantly stronger
482 relationships between job autonomy and dedication (z = 2.16) as
483 well as between dedication and psychological distress (z =2.41).
484 4. Discussion
485 The present study aimed to shed new light on the interplay
486 between burnout and work engagement by empirically investi-
487 gating whether both concepts are diametrical counterparts. More
488 specifically, this study simultaneously investigated the factorial
489 structure of the MBI-GS and UWES using ESEM, which allowed us
490 to examine the energy (emotional exhaustion-vigor) and identi-
491 fication (cynicism-dedication) continua. Our results offered little
492 support for both continua, revealing that work engagement dimen-
493 sions are highly intertwined and have stronger relationships
494 with each other than with their burnout counterpart. More-
495 over, integrating ESEM measurement models of burnout/work
496 engagement within a SEM that includes job characteristics and
497 health-related/motivational outcomes significantly alters the pat-
498 tern of results. As such, our results extend the understanding of the
499 health-impairment and motivational processes proposed by theJD-
500 R model. The theoretical implications of these results are discussed
501 below.
502 4.1. Theoretical contributions
503 4.1.1. The energy and identification continua
504 The energy and identification continua said to connect the core
505 dimensions of burnout and work engagement were investigated
506 by comparing two-factor and four-factor measurement models
507 (CFA and ESEM). Results offered little support for these continua
508 as the four-factor solutions fit the data significantly better than
509 the two-factor solutions. Furthermore, ESEM - which takes cross-
510 loadings of all indicators on all factors into account - allowed for
511 an in-depth examination of the energy and identification continua.
512 More specifically, we investigated whether strong cross-loadings
513 for the four dimensions were found from their theoretical coun-
514 terpart. The results did not follow such a pattern, suggesting that
515 burnout and work engagement are not conceptual opposites. With
516 regard to the two dimensions of burnout, results revealed no signif-
517 icant cross-loadings (.30 or higher) for any other items, suggesting
518 that both dimensions of burnout are distinct and tap into unique
519 work-related psychological experiences. These results support past
520 research validating the factorial structure of MBI-GS through CFA
521 (e.g., Hu & Schaufeli, 2011; Schutte, Toppinen, Kalimo, & Schaufeli,
522 2000), which showed that emotional exhaustion and cynicism are
523 best represented as separate factors (as opposed to a single fac-
524 tor). Overall, the results of the present study, in conjunction with
525 past CFA studies, highlight the relevance of investigating emotional
526 exhaustion and cynicism separately given that they reflect distinct
527 energetic and attitudinal experiences.
528 A different pattern of results was obtained for work engage-
529 ment. Results revealed several cross-loadings between the two
530 dimensions (using both the long and short versions of the UWES)
531 and none from items reflecting their burnout counterparts. The
532 vigor and dedication subscales of the UWES (and UWES-9) thus
represent similar psychological experiences that are difficult to 533
distinguish. This overlap is supported by the particularly strong 534
correlation found in this study between the two dimensions of 535
work engagement (.924 in the CFA measurement model and .597 in 536
the ESEM solution). This also concurs with past research showing 537
a strong interrelation between these dimensions, with correla- 538
tion coefficients usually exceeding .60 (e.g., Hallberg & Schaufeli, 539 2008; Schaufeli et al., 2002; Shimazu et al., 2008). Moreover, our Q8 540
results support those obtained in several validation studies (e.g., 541
Hallberg & Schaufeli, 2006; Shimazu et al., 2008) that found a 542
one-dimensional representation of work engagement to be equiv- 543
alent (or even superior) to a multi-dimensional representation (i.e., 544
vigor and dedication as distinct concepts). Overall, our results, like 545
previous CFA studies, highlight the considerable overlap between 546
vigor and dedication, suggesting that work engagement may be 547
best investigated as a one-dimensional construct. Specifically, our 548
results suggest that, because both dimensions are intertwined, 549
investigating work engagement globally (i.e., vigor and dedication 550
combined) as opposed to its two dimensions separately, may be 551
most parsimonious and appropriate. Further research is required 552
to pursue investigation of the uniqueness of vigor and dedication 553
from a conceptual standpoint (i.e., the fundamental nature of both 554
constructs). Researchers could also revisit the relationship between 555
vigor and dedication from a methodological standpoint by assess- 556
ing the adequacy of the UWES for assessing work engagement 557 (Schaufeli & Salanova, 2011). Q9 558
4.1.2. Health-impairment and motivational processes 559
Our study provides new insight into the relationships between 560
job characteristics (demands and resources), burnout/work 561
engagement and health-related as well as motivational outcomes. 562
Results reveal that SEM with CFA factors of burnout/work engage- 563
ment predominantly resulted in stronger relationships between 564
these factors and job characteristics. More specifically, half of the 565
links (five out of ten) were significantly stronger in the CFA solution 566
than in the ESEM solution. Of these five links, three were between 567
job resources (job autonomy, recognition) and work engage- 568
ment (the other significant links were between job demands and 569
cynicism/dedication). With regard to the relationships between 570
burnout/work engagement and indicators of employee functioning, 571
the ESEM solution revealed two cross-links (emotional exhaustion- 572
turnover intention and dedication-psychological distress) that 573
were not significant in the CFA solution. 574
Taken together, these findings reiterates the importance of 575
ESEM in structural solutions when investigating concepts that are 576
theoretically very closely related (Asparouhov & Muthen, 2009), 577
as it is the case for burnout and work engagement. Allowing 578
cross-loadings between these closely related concepts results in 579
a more adequate assessment of their interrelationship as well as 580
the association they have with their work-related antecedents 581
and outcomes (Asparouhov & Muthen, 2009). Our results show 582
that investigating the relationship between job characteristics and 583
burnout/work engagement through SEM with CFA factors may 584
result in an inflated representation ofthese relationships, especially 585
those involving job resources and work engagement (motivational 586
process). Moreover, these findings suggest that burnout and work 587
engagement may have more similar effects on employee func- 588
tioning than initially proposed by the JD-R model. That is, both 589
burnout and work engagement dimensions were found to pre- 590
dict health-related (i.e., psychological distress) and motivational 591
(i.e., turnover intention) outcomes. This corroborates findings of 592
past studies showing that the health-impairment and motivational 593
processes are interrelated (e.g., Bakker et al., 2003; Hakanen et al., 594
2008; Trepanier, Fernet, Austin, Forest, & Vallerand, 2014; Van den 595
Broeck, Vansteenkiste, De Witte, & Lens, 2008), as opposed to rela- 596
tively independent processes (Bakker & Demerouti, 2007). This also 597
llfil^^M ARTICLE IN PRESS
S.-G. Trépanier et al. / Burnout Research xxx (2015) xxx-xxx
598 underlines the relevance of investigating burnout/work engage-
599 ment simultaneously and of systematically examining cross-links
600 in order to more adequately capture the interplay between job
601 demands, job resources, burnout, work engagement, as well as
602 health-related and motivational outcomes (Schaufeli and Taris,
603 2014). This would result in a clearer representation of the poten-
604 tial motivational effects of job demands as well as the energetic
605 impact of job resources on employee functioning through burnout
606 and work engagement. Unfortunately, research to date has often
607 investigated both processes in isolation and has usually omitted
608 to evaluate all cross-links between these processes (e.g., Hakanen 60Q10 et al., 2006; Kounka et al., 2009; Schaufeli & Bakker, 2004).
610 4.2. Methodological implications
611 From a measurement standpoint, the results of this study offer
612 valuable insight to researchers and practitioners assessing burnout
613 and work engagement. By offering little support for the energy
614 and identification continua proposed to underlie the relationship
615 between the dimensions of burnout and work engagement, our
616 results show that these two concepts are distinct psychological
617 experiences and should be evaluated as such. Indeed, our findings
618 nuance past propositions suggesting that adding the scores of vigor
619 (dedication) items to the reversed scores of emotional exhaustion
620 (cynicism) items adequately represent employees' overall level of
621 "energy" and "identification" (e.g., Gonzalez-Roma et al., 2006).
622 Our results also call into question the assessment of burnout and
623 work engagement with the same instrument, as it is the case
624 with the OLBI (Demerouti & Bakker, 2008). The OLBI contains both
625 negatively and positively worded items said to reflect both ends
626 of the energy and identification continua. It has been proposed
627 that recoding positively framed items (reflecting vigor and dedi-
628 cation) results in the measurement of burnout whereas recoding
629 negatively framed items (reflecting exhaustion and disengage-
630 ment) results in the measurement of work engagement (Demerouti
631 & Bakker, 2008). However, our results indicate that high vigor
632 (dedication) does not necessarily imply low emotional exhaustion
633 (cynicism) and low cynicism (emotional exhaustion) does not nec-
634 essarily imply high dedication (vigor). Taken together, our results
635 suggest that both researchers and practitioners would benefit from
636 assessing burnout and work engagement (as well as their dimen-
637 sions) independently and through different instruments as they
638 reflect distinct concepts.
639 4.3. Limitations and conclusion
640 The present study has some limitations that should be
641 addressed. First, the fact that the study was conducted among
642 school teachers only raises concerns regarding the generalizabil-
643 ity of the findings to other working populations. Future research is
644 encouraged to validate our findings by investigating burnout and
645 work engagement, from both measurement and structural stand-
646 points, using ESEM in other occupations. Second, it is important to
647 note that in all tested models, particularly for the CFA measurement
648 models, the RMSEA fit value did not indicate a particularly good fit
649 (see Table 2), which hints at model misspecification (Hu & Bentler,
650 1999). Future research is needed in order to validate our measure-
651 ment and structural findings in other samples. Third, the structural
652 model tested in the present study focused on a limited num-
653 ber of job characteristics (i.e., work overload, job autonomy and
654 recognition) and only on negative employee functioning outcomes
655 (i.e., psychological distress and turnover intention). Moreover, it is
656 worth mentioning that as in previous studies (e.g., Fernet, Austin,
657 & Vallerand, 2012) the measure of job autonomy did not meet
658 the benchmark for reliability (coefficient H = .57). The revised AWS
659 (Leiter & Maslach, 2011), which comprises an additional item
to capture job autonomy, would certainly help represent more 660
adequately this construct. Future studies are encourage to repli- 661
cate the results using this job autonomy measure and other job 662
demands (e.g., role ambiguity, physical demands), job resources 663
(e.g., social support, skill utilization) as well as both positive and 664
negative indicators of employee functioning (e.g., in-role per- 665
formance, psychosomatic complaints, commitment). This should 666
provide additional support for the relevance of using ESEM analysis 667
when investigating the health-impairment and motivational pro- 668
cesses. Including objective and multi-source health-related (e.g., 669
sickness absence records) and motivational (e.g., supervisor ratings 670
of employee extra-role performance) indicators of employee func- 671
tioning would also strengthen the results obtained in the present 672
study, which relied solely on self-reported data. 673
In summary, although it has been proposed that burnout and 674
work engagement are conceptual counterparts, the results of this 675
study offer little support of this proposition. Our results also illus- 676
trate that ESEM represents a promising avenue for future burnout 677
and work engagement research as it may more adequately capture 678
the interplay between these concepts as well as their specific rela- 679
tionships with job characteristics and employee functioning (i.e., 680
health-impairment and motivational processes). 681
Conflict of interest Q11682
The authors declare that there are no conflicts of interest. 683
Uncited references Q12684
Bakker, Albrecht and Leiter (2011), Demerouti, Bakker, 685
Vardakou and Kantas (2003), Marsh et al. (2013), Schumacker and 686
Lomax (1996), Seppala et al. (2009) and Shirom and Melamed 687
(2006). 688
References 689
Asparouhov, T., & Muthen, B. (2009). Exploratory structural equation modeling. 690
Structural Equation Modeling, 16,397-438. 691
Bakker, A. B., Albrecht, S. L., & Leiter, M. P. (2011). Key questions regarding work 692
engagement. European Journal of Work and Organizational Psychology, 20(1), 693
4-28. 694
Bakker, A. B., & Demerouti, E. (2007). The job demands-resources model: State of 695
the art. Journal ofManagerial Psychology, 22, 309-328. 696
Bakker, A. B., Demerouti, E., & Schaufeli, W. B. (2003). Dual processes at work in a call 697
center: An application of the job demands-resources model. European Journal of 698
Work and Organizational Psychology, 12(4), 393-417. 699
Bakker, A. B., Demerouti, E., & Sanz-Vergel, A. I. (2014). Burnout and work engage- 700
ment: The JD-R approach. Annual Review of Organizational Psychology and 701
Organizational Behavior, 1(1), 389-411. 702
Bakker, A. B., Schaufeli, W. B., Leiter, M. P., & Taris, T. W. (2008). Work engage- 703
ment: An emerging concept in occupational health psychology. Work & Stress, 704
22,187-200. 705
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). New 706
Jersey: Lawrence Erlbaum. 707
Crawford, E. R., LePine, J. A., & Rich, B. L. (2010). Linking job demands and resources to 708
employee engagement and burnout: A theoretical extension and meta-analytic 709
test. Journal of Applied Psychology, 95(5), 834-848. 710
Demerouti, E., & Bakker, A. B. (2008). The oldenburg burnout inventory: A good 711
alternative to measure burnout and engagement. In J. Halbesleben (Ed.), Stress 712
and burnout in health care (pp. 65-78). Hauppage, NY: Nova Sciences. 713
Demerouti, E., Bakker, A. B., Nachreiner, F., & Schaufeli, W. B. (2001). The job 714
demands-resources model of burnout. Journal of Applied Psychology, 86(3), 715
499-512. 716
Demerouti, E., Bakker, A. B., Vardakou, I., & Kantas, A. (2003). The convergent valid- 717
ity of two burnout instruments: A multitrait-multimethod analysis. European 718
Journal ofPsychological Assessment, 19, 12-23. 719
Demerouti, E., Mostert, K., & Bakker, A. B. (2010). Burnout and work engagement: 720
A thorough investigation of the independency of both constructs. Journal of 721
Occupational Health Psychology, 15(3), 209-222. 722
Fernet, C., Austin, S., & Vallerand, R. J. (2012). The effects of work motivation on 723
employee exhaustion and commitment: An extension of the JD-R model. Work 724
& Stress, 26,213-229. 725
Gonzalez-Roma, V., Schaufeli, W. B., Bakker, A. B., & Lloret, S. (2006). Burnout and 726
work engagement: Independent factors or opposite poles? Journal of Vocational 727
Behavior, 68,165-174. 728
lllli^^^M Al IULE IN PRESS
S.-G. Trépanier et al. / Burnout Research xxx (2015) xxx-xxx 9
Guay, F., Morin, A. J. S., Litalien, D., Valois, P., & Vallerand, R. J. (2015). Application of exploratory structural equation modeling to evaluate the academic motivation scale. Journal of Experimental Education, 83(1), 51-82.
Hakanen, J. J., Bakker, A. B., & Schaufeli, W. B. (2006). Burnout and work engagement among teachers. Journal of School Psychology, 43, 495-513.
Hakanen, J. J., Schaufeli, W. B., & Ahola, K. (2008). The job demands-resources model: A three year cross-lagged study of burnout, depression, commitment and work engagement. Work & Stress, 22(3), 224-241.
Halbesleben, J. R. B. (2010). A meta-analysis of work engagement: Relationships with burnout, demands, resources, and consequences. In A. B. Bakker, & M. P. Leiter (Eds.), Work engagement: A handbook of essential theory and research (pp. 102-117). New York, NY: Psychology Press.
Hallberg, U., & Schaufeli, W. B. (2006). Same same" but different: Can work engagement be discriminated from job involvement and organizational commitment? European Journal of Psychology, 11,119-127.
Hancock, G. R., & Mueller, R. O. (2001). Rethinking construct reliability within latent variable systems. In R. Cudeck, S. du Toit, & D. Sorbom (Eds.), Structural equation modeling: Present and future - A festschrift in honor of Karl Joreskog (pp. 195-216). Lincolnwood, IL: Scientific Software International.
Hooper, D., Coughlan, J., & Mullen, M. (2008). Structural equation modelling: Guidelines for determining model fit. Electronic Journal of Business Research Methods, 6(1), 53-60.
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.
Hu, Q., & Schaufeli, W. B. (2011). The convergent validity of four burnout measures in a Chinese Sample: A confirmatory factor-analytic approach. Applied Psychology: An International Review, 61(1), 87-111.
Hu, Q., Schaufeli, W. B., & Taris, T. W. (2011). The job demands-resources model: An analysis of additive and joint effects of demands and resources. Journal of Vocational Behavior, 79,181-190.
Ilfeld, F. W. (1976). Further validation of a psychiatric symptom index in a normal population. Psychological Reports, 39,1215-1228.
Korunka, C., Kubicek, B., Schaufeli, W., & Hoonakker, P. L. T. (2009). Work engagement and burnout: Testing the robustness ofthe job demands-resources model. Journal of Positive Psychology, 4(3), 243-255.
Leiter, M. P., & Bakker, A. B. (2010). Work engagement: Introduction. In A. B. Bakker, & M. P. Leiter (Eds.), Work engagement: Ahandbook of essential theory and research (pp. 1-9). New York, NY: Psychology Press.
Leiter, M. P., & Maslach, C. (2004). Areas of worklife: A structured approach to organizational predictors of job burnout. In P. Perrewe, & D. C. Ganster (Eds.), Research in occupational stress and well-being: Emotional and physiological processes and positive intervention strategies (vol. 3) (pp. 91-134). Oxford, UK: JAI Press/Elsevier.
Leiter, M. P., & Maslach, C. (2011). Areas of worklife scale manual (5th ed.). Palo Alto, CA: Mindgarden Publishing.
Marsh, H. W., Liem, G. A. D., Martin, A. J., Morin, A. J. S., & Nagengast, B. (2011). Methodological-measurement fruitfulness of exploratory structural equation modeling (ESEM): New approaches to key substantive issues in motivation and engagement. Journal of Psychoeducational Assessment, 29,322-346.
Marsh, H. W., Muthen, B., Asparouhov, T., Ludtke, O., Robitzsch, A., Morin, A. J. S., et al. (2009). Exploratory structural equation modeling, integrating CFA and EFA: Application to students' evaluations of university teaching. Structural Equation Modeling, 16, 439-476.
Marsh, H. W., Nagengast, B., & Morin, A. J. S. (2012). Measurement invariance of Big-Five factors over the life span: ESEM tests of gender, age, plasticity, maturity, and La Dolce Vita effects. Developmental Psychology, 49(6), 1194-1218.
Marsh, H. W., Vallerand, R. J., Lafreniere, M.-A. K., Parker, P., Morin, A. J. S., Car-bonneau, N., et al. (2013). Passion: does one scale fit all. Construct validity of two-factor passion scale and psychometric invariance over different activities and languages. Psychological Assessment, 25(3), 796-809.
Maslach, C. (1982). Understanding burnout: Definitional issues in analyzing a complex phenomenon. In W. S. Paine (Ed.), Job stress and burnout (pp. 29-40). Beverly Hills, CA: Sage.
Morin, A. J. S., Marsh, H. W., & Nagengast, B. (2013). Exploratory Structural Equation Modeling. In G. R. Hancock, & R. O. Mueller (Eds.), Structural equation modeling: A second course. Charlotte, NC: Information Age Publishing, Inc.
Muthén, L. K., & Muthén, B. O. (2012). Mplus user's guide (6th ed.). Los Angeles, CA: Muthén & Muthén.
O'Driscoll, M. P., & Beehr, T. A. (1994). Supervisor behaviors, role stressors and uncertainty as predictors of personal outcomes for subordinates. Journal of Organizational Behavior, 15(2), 141-155.
Préville, M., Boyer, R., Potvin, L., Perrault, C., & Légaré, G. (1992). La détresse psychologique: Détermination de la fiabilité et de la validité de la mesure utilisée dans l'enquête Santé Québec [Psychological distress: Determining the reliability and validity of the measure used in the Quebec Health Survey]. Retrieved from Institut de la statistique Québec http://www.stat.gouv.qc.ca/publications/sante/pubs_ anterieur/detresse _psychologique1987.pdf
Richardsen, A. M., Burke, R. J., & Martinussen, M. (2006). Work and health outcomes among police officers: The mediating role of police cynicism and engagement. International Journal of Stress Management, 13(4), 555-574.
Schaufeli, W. B., & Bakker, A. B. (2004). Job demands, job resources, and their relationship with burnout and engagement: a multi-sample study. Journal of Organizational Behavior, 25, 293-315.
Schaufeli, W. B., Bakker, A. B., & Salanova, M. (2006). The measurement of work engagement with a short questionnaire: A cross-national study. Educational and Psychological Measurement, 66, 701-716.
Schaufeli, W. B., Leiter, M. P., Maslach, C., & Jackson, S. E. (1996). Maslach burnout inventory - general survey. In C. Maslach, S. E. Jackson, & M. P. Leiter(Eds.), The Maslach burnout inventory-test manual (3rd ed., vol. 3, pp. 22-26). Palo Alto, CA: Consulting Psychologists Press.
Schaufeli, W. B., Salanova, M., Gonzâlez-Româ, V., & Bakker, A. B. (2002). The measurement of engagement and burnout: A two sample confirmatory factor analytic approach. Journal of Happiness Studies, 3, 71-92.
Schaufeli, W. B., & Taris, T. W. (2005). The conceptualization and measurement of burnout: Common ground and worlds apart. Work & Stress, 19(3), 256-262.
Schumacker, R. E., & Lomax, R. G. (1996). A beginner's guide to structural equation modeling. Mahwah, NJ: Lawrence Erlbaum Associates.
Schutte, N., Toppinen, S., Kalimo, R., & Schaufeli, W. B. (2000). The factorial validity of the Maslach Burnout Inventory-General Survey (MBI-GS) across nations and occupations. Journal of Occupational and Organizational Psychology, 73, 53-66.
Seppälä, P., Mauno, S., Feldt, T., Hakanen, J., Kinnunen, U., Tolvanen, A., et al. (2009). The construct validity of the Utrecht work engagement scale: Multisample and longitudinal evidence. Journal of Happiness Studies, 10,459-481.
Shirom, A., & Melamed, S. (2006). A comparison of the construct validity of two burnout measures in two groups of professionals. International Journal of Stress Management, 13(2), 176-200.
Shimazu, A., Schaufeli, W. B., Nashiwa, H., Kato, A., Sakamoto, M., Irimajiri, H., et al. (2008). Work engagement in Japan, validation of the Japanese Version of the Utrecht Work Engagement Scale. Applied Psychology: An International Review, 37(3), 510-523.
Steiger, J. H. (2007). Understanding the limitations ofglobal fit assessment in structural equation modeling. Personality and Individual Differences, 42(5), 893-898.
Tabachnick, B. G., & Fidell, L. S. (2007). Using multivariate analysis. Boston: Pearson Education.
Trépanier, S.-G., Fernet, C., Austin, S., Forest, J., & Vallerand, R. J. (2014). Linking job demands and resources to burnout and work engagement: Does passion underlie these differential relationships? Motivation and Emotion, 38(3), 353-366.
Van den Broeck, A., Vansteenkiste, M., De Witte, H., & Lens, W. (2008). Explaining the relationships between job characteristics, burnout, and engagement: The role of basic psychological need satisfaction. Work FsStress, 22, 277-294.
