Journal of Occupational Health Psychology 2010, Vol. 15, No.

3, 209 222

2010 American Psychological Association 1076-8998/10/$12.00 DOI: 10.1037/a0019408

Burnout and Work Engagement: A Thorough Investigation of the Independency of Both Constructs
Evangelia Demerouti
Utrecht University, and Eindhoven University of Technology

Karina Mostert
North-West University

Arnold B. Bakker
Erasmus University Rotterdam
This study among 528 South African employees working in the construction industry examined the dimensionality of burnout and work engagement, using the Maslach Burnout InventoryGeneral Survey, the Oldenburg Burnout Inventory, and the Utrecht Work Engagement Scale. On the basis of the literature, we predicted that cynicism and dedication are opposite ends of one underlying attitude dimension (called identication), and that exhaustion and vigor are opposite ends of one energy dimension. Conrmatory factor analyses showed that while the attitude constructs represent opposite ends of one continuum, the energy constructs do notalthough they are highly correlated. These ndings are also supported by the pattern of relationships between burnout and work engagement on the one hand, and predictors (i.e., work pressure, autonomy) and outcomes (i.e., organizational commitment, mental health) on the other hand. Implications for the measurement and conceptualization of burnout and work engagement are discussed. Keywords: burnout, conrmatory factor analysis, dimensionality, work engagement

Most scholars agree that burned-out employees are characterized by high levels of exhaustion and negative attitudes toward their work (cynicism; Maslach, Schaufeli & Leiter, 2001). Recently, Schaufeli and Bakker (2003, 2004; Schaufeli, Salanova, Gonzalez Roma, & Bakker, 2002) introduced work engage ment as the hypothetical antipode of burnout. Accordingly, engaged employees are characterized by high levels of energy and dedication to their work. One unclear issue is whether the dimensions of burnout and work engagement are each others opposite, which would mean that one instrument (covering both ends of the continuum) would be sufcient to measure both constructs. Demerouti, Bakker, Varda-

Evangelia Demerouti, Department of Social and Organizational Psychology, Utrecht University, and Department of Technology Management, Human Performance Management Group, Eindhoven University of Technology; Karina Mostert, School of Human Resource Management, NorthWest University; and Arnold B. Bakker, Department of Work and Organizational Psychology, Erasmus University Rotterdam. Correspondence concerning this article should be addressed to Evangelia Demerouti, Eindhoven University of Technology, Industrial Engineering & Innovation Sciences, Human Performance Management Group, P.O. Box 513, 5600 MB Eindhoven, the Netherlands. E-mail: [email protected]

kou, and Kantas (2003) developed the Oldenburg Burnout Inventory (OLBI) which contains questions on both ends of the exhaustion-vigor and cynicismdedication continua, hereafter referred to as energy and identication dimensions (see also Gonzalez Roma, Schaufeli, Bakker & Lloret, 2006).1 The present study builds on the study of Gonzalez Roma et al. (2006), and adds to the literature in several ways. First, we will use a parametric scaling technique namely conrmatory factor analysis to test the dimensionality of the energy and identication dimension of burnout and of work engagement. This will overcome an important drawback of the MSPprogram used by Gonzalez-Roma et al. to conduct Mokken analysisthat is, that the sequential item selection and scale construction procedure may not nd the dominant underlying dimensionality of the responses to a set of items (Van Abswoude, Vermunt, Hemker, & Van der Ark, 2004). Moreover, Mokken analysis can be applied to scales including items with
1 In the following, we will use the term identication to describe the hypothetical dimension running from distancing [cynicism (MBI-GS) or disengagement (OLBI)] to dedication (UWES and OLBI). In addition, we will use the term energy to describe the hypothetical dimension running from exhaustion (OLBI and MBI-GS) to vigor (UWES and OLBI).

209

210

DEMEROUTI, MOSTERT, AND BAKKER

a hierarchical property that is, that can be ordered by degree of difculty. However, none of the instruments used in this study are known for including hierarchical structured items. Second, in addition to the Maslach Burnout Inventory-General Survey (MBI-GS; Maslach, Jackson, & Leiter, 1996) and the Utrecht Work Engagement Scale (UWES; Schaufeli et al., 2002) we will use the OLBI, which is a valid instrument that can be used to measure the energy and identication dimensions of burnout and work engagement simultaneously as bipolar constructs. We focused on these instruments because they include both core dimensions of burnout and work engagement, namely a vigor/exhaustion dimension and an identication/distancing dimension, while instruments like the Shirom-Melamed Burnout Scale (Shirom, 2003) or the Copenhagen Burnout Inventory (Kristensen, Borritz, Villadsen & Christensen, 2005) focus solely on vigor/exhaustion. Third, next to the factor structure we will also examine the pattern of relationships of the (bipolar and/or unipolar) dimensions of burnout and work engagement with relevant job characteristics (work pressure and job autonomy) and organizational outcomes (organizational commitment and mental health). We focus on these constructs because they have been studied most often as being related to the energy or identication dimensions of burnout or engagement.

Measurement of Burnout and Work Engagement

The most commonly used instrument for the measurement of burnout is the MBI-GS (Schaufeli, Leiter, Maslach, & Jackson, 1996). Based on the notion that emotional exhaustion, depersonalization and personal accomplishment (representing symptoms of burnout specic for human services) can be broadened beyond the interpersonal domain that is characteristic for the human services, they distinguished three generic burnout dimensions that were labeled exhaustion, cynicism and professional efcacy, respectively. Many empirical ndings point to the central role of exhaustion and cynicism as the core dimensions of burnout, as opposed to the third componentlack of professional efcacy (Lee & Ashforth, 1996). As a result, the third dimension measured with the MBI-GS was excluded from this study. Several studies have supported the invariance of the MBI-GS factor structure across various occupational groups (Bakker, Demerouti & Schaufeli, 2002; Leiter & Schaufeli, 1996), and across nations

(Richardsen & Martinussen, 2004; Schutte, Toppinen, Kalimo, & Schaufeli, 2000). Unfortunately, the MBI-GS has one important psychometric shortcoming, namely that the items within each subscale are all framed in the same direction. Accordingly, all exhaustion and cynicism items are phrased negatively, whereas all professional efcacy items are phrased positively. From a psychometric point of view, such one-sided scales are inferior to scales that include both positively and negatively worded items (Price & Mueller, 1986) because they can lead to articial factor solutions in which positively and negatively worded items are likely to cluster (Demerouti & Nachreiner, 1996; cf. Doty & Glick, 1998) or may show articial relationships with other constructs (Lee & Ashforth, 1990). The UWES (Schaufeli & Bakker, 2003, 2010; Schaufeli et al., 2002) has been developed to measure work engagement dened as a positive, fullling, work-related state of mind that is characterized by vigor, dedication, and absorption. Vigor refers to high levels of energy and mental resilience while working. Dedication refers to a sense of signicance, enthusiasm, inspiration, and pride. Vigor and dedication are the direct positive opposites of exhaustion and cynicism, respectively. Absorption is excluded from the present study because burnout does not contain any parallel dimension to this dimension. The UWES has been validated in several countries (e.g., Schaufeli et al., 2002; Schaufeli & Bakker, 2004; Storm & Rothmann, 2003; Yi-Wen & Yi-Qun, 2005). However, some studies found a one- instead of a three-factor structure of work engagement (e.g., Sonnentag, 2003). We propose an alternative measure of burnout and work engagement: The OLdenburg Burnout Inventory (OLBI; Demerouti, 1999; Demerouti & Nachreiner, 1998). It includes positively and negatively framed items to assess the two core dimensions of burnout: exhaustion and disengagement from work. Exhaustion is dened as a consequence of intensive physical, affective and cognitive strain, that is, as a long-term consequence of prolonged exposure to certain job demands. Contrary to exhaustion as operationalized in the MBI-GS, the OLBI covers affective but also physical and cognitive aspects of exhaustion. Such an operationalization of exhaustion/vigor covers more thoroughly peoples intrinsic energetic resources, that is, emotional robustness, cognitive liveliness and physical vigor (Shirom, 2003) and enables the application of the instrument to those workers with physical and cognitive work. Disengagement refers to distancing oneself from ones work in gen-

BURNOUT AND WORK ENGAGEMENT

211

eral, work object, and work content. Moreover, the disengagement-items concern the relationship between employees and their jobs, particularly with respect to identication with work and willingness to continue in the same occupation. Depersonalization is consequently only one form of disengagement which is directed toward customers. The factorial validity of the OLBI has been conrmed in studies conducted in different countries (Demerouti & Bakker, 2008; Demerouti, Bakker, Nachreiner, & Ebbinghaus, 2002; Halbesleben & Demerouti, 2005; Demerouti et al., 2003). Following a multitrait multimethod approach, Demerouti et al. (2003) and Halbesleben and Demerouti (2005) conrmed the convergent validity of the OLBI and MBI-GS.

The Dimensionality of Burnout and Work Engagement

There are different views regarding the dimensionality of burnout and work engagement. Demerouti and colleagues (2001, 2003) assume that the dimensions of burnout and work engagement are bipolar dimensions. This is reected in the OLBI which includes both negatively and positively worded items so that both ends of the continuum are measured. In other words, the exhaustion and disengagement subscales include items that refer to their opposites, namely vigor and dedication, respectively. Positively framed items should be reverse-coded if one wants to assess burnout. Alternatively, to assess work engagement the negatively framed items should be recoded (Demerouti & Bakker, 2008). Maslach and Leiter (1997) agree with this standpoint. They rephrased burnout as an erosion of engagement with the job, whereby energy turns into exhaustion, involvement turns into cynicism, and efcacy turns into ineffectiveness. In their view, work engagement is characterized by energy, involvement and professional efcacy, which are the direct (perfectly inversely related) opposites of the three burnout dimensions. However, it should be noted that their MBI-GS includes negative items only. Therefore, low scores on exhaustion and cynicism cannot be taken as being representative of vigor and dedication, since employees who indicate that they are not fatigued are not necessarily full of energy. Schaufeli and Bakker (2003, 2010; Schaufeli et al., 2002) argue that work engagement cannot be measured by the opposite prole of the MBI-GS, because, even though in conceptual terms engagement is the positive antithesis of burnout, the content and

consequently the measurement of both concepts is different. As the MBI-GS includes only negatively worded items, it is difcult to conclude that individuals who reject a negatively worded statement would automatically agree with a positively worded one. Thus, Schaufeli and Bakker (2003) proposed that burnout and work engagement should be conceived as two opposite concepts that should be measured independently with different instruments. A direct test of the dimensionality of burnout and work engagement has been conducted by Gonzalez Roma et al. (2006). They used the MBI-GS and the UWES to test the hypothesis that items reecting exhaustion-vigor and cynicism-dedication are scalable on two distinct underlying bipolar dimensions (labeled energy and identication, respectively). Using a nonparametric scaling technique, they showed that these core burnout and engagement dimensions can indeed be seen as opposites of each other along two distinct bipolar dimensions (energy vs. identication). However, a closer look at their ndings reveals that the exhaustionvigor items constitute a weak to moderate energy dimension, and that the cynicism dedication items constitute a strong identication dimension. Nevertheless, it can be concluded from this study that negatively and positively framed items can be used to assess the core dimensions of burnout and work engagement. Specically: Hypothesis 1: Disengagement/cynicism and dedication are opposite ends of one dimension. Hypothesis 2: Exhaustion and vigor are opposite ends of one dimension. Work pressure and autonomy are two job characteristics that have been related to burnout and to work engagement. Specically, work pressure has the strongest positive relationship with exhaustion (Demerouti, Bakker, & Bulters, 2004; Lewig, Xanthopoulou, Bakker, Dollard, & Metzer, 2007; Hakanen, Bakker, & Schaufeli, 2006; Rothmann & Pieterse, 2007); and a less strong but negative relationship with vigor (Hakanen et al., 2006; Rothmann & Pieterse, 2007). However, some authors found a nonsignicant relationship between work/time pressure and exhaustion or vigor (Mauno, Kinnunen, & Ruokolainen, 2007). The relationship between work pressure and the identication components of burnout and work engagement is, however, weak (e.g., Bakker, Demerouti, & Verbeke, 2004; Hakanen et al., 2006; Rothmann & Pieterse, 2007). Autonomy seems to be related to the identication dimensions and the energy dimensions

212

DEMEROUTI, MOSTERT, AND BAKKER

(Bakker et al., 2004; Demerouti et al., 2001; Schaufeli & Bakker, 2004). It shows a negative relationship with exhaustion and cynicism (Bakker et al., 2004; Hakanen et al., 2006; Koekemoer & Mostert, 2006; Schaufeli & Bakker, 2004), and a positive relationship with vigor and dedication (Hakanen et al., 2006; Mauno et al., 2007). Organizational commitment is an outcome that is particularly related to the identication components of burnout and work engagement (Schaufeli & Bakker, 2004) and weakly related/unrelated to the energy components, specically exhaustion (Hakanen et al., 2006; Llorens, Bakker, Schaufeli, & Salonova, 2006). Finally, mental health shows a stronger relationship with the energy dimensions and in particular with exhaustion (Hakanen et al., 2006; Jackson & Rothmann, 2005; Lewig et al., 2007; Schaufeli & Bakker, 2004). What is of interest for our research question is whether the two ends of the energy and identication dimensions show the same pattern of relationships with these constructs. Similar relationships would be evidence for bipolar constructs, while differential relations would substantiate the argument for independent (unipolar) dimensions. For instance, if exhaustion and vigor are equally strong related to work pressure (but in the opposite direction) this would suggest that they represent opposite poles of one dimension. If, however, one of them is substantially stronger related to work pressure this would mean that they represent different and thus independent dimensions. Because there is no clear evidence for differential relationships between these constructs and the two ends of the energy and identication dimensions we formulated the following hypotheses: Hypothesis 3: Cynicism/disengagement and dedication/engagement will be equally strong related to other constructs (work pressure, autonomy, organizational commitment, mental health), but in the opposite direction. Hypothesis 4: Exhaustion and vigor will be equally strong related to other constructs (work pressure, autonomy, organizational commitment, mental health), but in the opposite direction.

The response rate was 53%. After permission was obtained from executive management, the managers, Human Resources department, and employee/ employer committees were informed of the study during management meetings. Thereafter, all employees received paper-and-pencil questionnaires and envelopes at their work that could be returned to the researchers involved. A letter explaining the purpose of the research accompanied the questionnaire. The employees were kindly requested to ll in the questionnaire in private and send it to the Human Resources department, where the researchers collected all the completed questionnaires. Participation was voluntary, and the condentiality and anonymity of the answers was emphasized. The majority of the participants worked in the Construction (40.2%) and Mining (24.2%) units, while the rest worked in the Shared Services (12.3%), Handling (8.3%), Energy (5.1%), Rental (5.1%), and Agriculture (4.8%) departments. The participants were predominantly male (71.5%), while 62.7% were White, 20.4% were African, 11.2% were Colored, and 3.1% were Indian. The mean age was 39.61 (SD 11.02). A total of 58.5% of the participants had a high school qualication (Grade 10-Grade 12), while 39.2% possessed a (technical college) diploma or university degree. Most participants were married/ living with a partner, with children living at home (50.6%).

Instruments2
MBI-GS. We used the MBI-GS (Schaufeli et al., 1996) to assess the core burnout dimensions with two subscales, namely Exhaustion and Cynicism. Exhaustion was measured with ve items (e.g., I feel emotionally drained from my work). Cynicism was assessed with ve items (e.g., I have become less enthusiastic about my work). All items are scored on a seven-point scale, ranging from (0) never to (6) every day. High scores on exhaustion and cynicism indicate burnout. OLBI. The OLBI originally distinguishes an exhaustion and disengagement dimension. However, both subscales include four items that are positively worded and four items that are negatively worded. This means that both ends of the energy and identication dimensions are included in the OLBI. The
2 Le Roux (2005) and Rost (2007) have conrmed the construct equivalence of the instruments used in the present study for different language and educational groups.

Method Participants and Procedure

A cross-sectional survey was conducted using a convenience sample of employees of a company in the South African construction industry (N 528).

BURNOUT AND WORK ENGAGEMENT

213

answering categories are (1) strongly agree to (4) strongly disagree. The OLBI items are displayed in the Appendix. UWES. The UWES (Schaufeli & Bakker, 2003; Schaufeli et al., 2002) was used to assess the two core dimensions of work engagement, namely vigor and dedication. Vigor was assessed with six items (e.g., At my work, I feel bursting with energy). Dedication was assessed with ve items (e.g., I nd the work that I do full of purpose and meaning). All items are scored on a seven-point rating scale, ranging from (0) never to (6) every day. High scores indicate work engagement. Work pressure was measured with six items that were adapted from the Job Content Questionnaire (Karasek, 1985). The original statements were rephrased as questions (e.g., Are you asked to do an excessive amount of work?). Items were scored on a scale ranging from (1) almost never to (4) always, with higher scores indicating higher job pressure. Autonomy was measured with six items from the validated questionnaire of Van Veldhoven, Meijman, Broersen and Fortuin (1997) (e.g., Can you decide for yourself how to carry out your work?). Items were scored on a four-point rating scale: (1) almost never to (4) always. Higher scores signify a higher level of autonomy. Mental health was measured with the General Health Questionnaire (GHQ-28, Goldberg & Williams, 1988). The GHQ-28 is a 28 item questionnaire generally used for the screening of mental illness. The GHQ-28 asks participants to report if they have had any medical complaints and how their general health had been over the past few weeks, rating them on a 4-point scale ranging from (1) better than usual to 4 much worse than usual. The scale taps four factors: somatic symptoms, anxiety and insomnia, social dysfunction and depression. Scores were coded such that higher overall scores indicate better mental health. Organizational commitment was measured with ve items of the affective organizational commitment scale developed by Meyer, Allen, and Smith (1993). An example item is This organization has a great deal of personal meaning for me. Items were rated on a 5-point scale ranging from (1) totally agree to (5) totally disagree.

icism for the MBI-GS, exhaustion and disengagement for the OLBI, vigor and dedication for the UWES) tted responses to all instruments substantially better than did one-factor solutions. All items had signicant loadings on the expected factors except for the third item of the cynicism scale (i.e., I just want to do my work and not be bothered). This is consistent with earlier studies (Schutte et al., 2000; Storm & Rothmann, 2003). Consequently, we decided not to include this item in further analyses. We tted the responses to all three instruments simultaneously to the data. However, the energy dimensions (OLBI-exhaustion, OLBI-vigor, MBI-exhaustion, and UWES-vigor) were analyzed separately from the identication dimensions (OLBI-disengagement, OLBI-dedication, MBI-cynicism, and UWES-dedication). This was done in order to avoid building large models (in this case including 36 manifest variables) that generally show a poor t to the data. Bentler and Chou (1987) suggest that models should not exceed the total of 20 manifest variables because in large models with large sample sizes the sample size multiplier that transforms the t function into a 2-variate will multiply a small lack of t into a large statistic (p. 97). Building smaller models still allows testing our hypotheses. We followed the same way of modeling to test the relationships between the energy and identication dimensions with other variables (work pressure, autonomy, mental health, and commitment) save one difference: we included age and gender as control variables. Specically, age and gender had a path to all manifest variables of the models. CFAs were conducted with AMOS 7 (Arbuckle, 2006). Next to the inspection of the goodness-of-t indices we performed chi-square difference tests in order to compare alternative, nested models.

Results
Cronbachs alpha and bivariate correlations between the study variables are displayed in Table 1. Note that while all (sub-)scales had sufcient reliability, for OLBI vigor this was .63. However, we had to keep this subscale in order to retain a minimum of two indicators for each end of the continua.

Statistical Analysis
In preliminary, unreported CFAs, one- and twofactor models were tted to responses to each of the three instruments separately. The results indicated that two-factor model solutions (exhaustion and cyn-

Inferring Identication and Energy Dimensions

The dimensionality of the identication dimension was tested with alternative models (see Figure 1). We tested whether considering separate identication factors that is, MBI-cynicism, UWES-dedication,

214

DEMEROUTI, MOSTERT, AND BAKKER

Table 1 Means, Standard Deviations, and Bivariate Correlations of the Study Variables
Mean SD 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. MBI exhaustion MBI cynicism UWES vigor UWES dedication OLBI exhaustion total1 OLBI exhaustion OLBI vigor OLBI disengagement total1 OLBI disengagement OLBI dedication Mental health Work pressure Autonomy Organizational commitment 2.51 2.20 4.48 4.95 2.17 2.47 1.87 2.07 1.97 2.85 .68 2.26 2.30 2.04 1.40 1.26 1.10 1.19 .57 .74 .57 .55 .62 .69 .45 .55 .63 .79 1 .82 .44 .35 .42 .62 .60 .45 .52 .38 .49 .54 .27 .28 .31 2 .73 .42 .48 .45 .41 .37 .54 .37 .54 .39 .08 .18 .37 3 4 5 6 7 8 9 10 11 12 13 14

.69 .71 .53 .44 .48 .55 .48 .45 .37 .09 .32 .36

.85 .50 .42 .45 .68 .65 .49 .41 .02 .36 .50 .74 .90 .82 .67 .50 .62 .61 .13 .34 .30 .78 .49 .62 .37 .65 .56 .17 .28 .26

.63 .53 .51 .39 .49 .04 .31 .27 .79 .82 .85 .49 .04 .36 .48 .69 .41 .39 .04 .33 .49

.71 .43 .11 .28 .32

.94 .16 .31 .19

.77 .09 .78 .04 .22 .87

Note. Cronbachs alpha on the diagonal, N 528. 1 OLBI exhaustion total and OLBI disengagement total refer to the average score of all positively and negatively worded items of the original exhaustion and disengagement OLBI dimensions, respectively. All correlations r |.13| are signicant at p .01, while correlations |.09| r |.13| are signicant at p .05.

OLBI-disengagement, and OLBI-dedication (Model 1), was better compared to two second-order factors of distancing and dedication (Model 3) or compared to only one second-order factor of identication (Model 2). Note that OLBI-disengagement included the four negatively formulated items and OLBIdedication the four positively formulated items of the disengagement scale. In this way, we had two indicators for each end of the continuum (i.e., two scales for distancing and two for dedication), which is useful for building second-order latent factors. The same procedure was followed for the energy dimensions in a separate series of analyses including MBIexhaustion, UWES-vigor, OLBI-exhaustion and OLBI-vigor, OLBI-exhaustion, and OLBI-vigor as rst-order factors. Model 1 explains responses to the items in terms of four rst-order factors. This rst-order model is important because its t establishes an upper limit for the higher-order models (cf. Marsh, Antill, & Cunningham, 1989). As can be seen in Table 2, the t of Model 1 is reasonable for both the identication and the energy dimensions. For both dimensions, the factor structure is well-dened in that all factor loadings were statistically signicant and each of the four factors accounts for a signicant portion of the variance. The aim of the higher-order models is to describe correlations among rst-order factors in terms of

higher-order factors. Three different second-order models were tested. In Model 2, the four rst-order factors of the identication dimensions were used to dene an overall identication factor (assuming a bipolar dimension). In Model 3, MBI-cynicism and OLBI-disengagement loaded on a distancing secondorder factor, while UWES-dedication and OLBIdedication loaded on a dedication second-order factor (assuming a unipolar dimension). The second-order factors were allowed to correlate. In Model 4 we tested the discriminant validity of the two secondorder latent factors (of Model 3) by constraining their correlation to be 1 (implying identical constructs, cf. Bagozzi, 1993). Following the same logic we tested parallel models for the energy dimensions using the respective four rst-order factors. All models were nested in Model 1 so that none can t the data better than the rst-order factors model but they were more parsimonious in that they included fewer parameters. First, we discuss results regarding the identication dimensions. Model 2, including a single higherorder factor, tted the data signicantly worse than Model 1. This means that much of the variation among the rst-order factors is unexplained by a global identication factor. Model 3 (positing two higher-order factors) provides a better t to the data than the one-factor model (Model 2) and is not signicantly worse than the rst-order factors model

BURNOUT AND WORK ENGAGEMENT

1 1

215

e4 e3 e2 e1

1 1 1

CY1 CY2 CY4 CY5

1

DE1 DE2 DE3 DE4 DE5

1 1 1 1

e13 e14 e15 e16 e17

MBI Cynicism

UWES Dedication

e8

OLBI_3D OLBI_6D OLBI_9D OLBI_11D

1

e7 1 e6
1

OLBI Disengagement

OLBI Dedication

OLBI_1D OLBI_7D OLBI_13D OLBI_15D

1 1 1

e9 e10 e11 e12

e5 Model 1

e18
1 1

e19
1 1 1

e4 e3 e2 e1

1 1 1

CY1 CY2 CY4 CY5

1

DE1 DE2 DE3 DE4 DE5

1 1 1 1

e13 e14 e15 e16 e17

MBI Cynicism
1

UWES Dedication

Identification
1 1

e8

OLBI_3D OLBI_6D OLBI_9D OLBI_11D

1

e7 1 e6
1

OLBI Disengagement
1

OLBI Dedication
1

OLBI_1D OLBI_7D OLBI_13D OLBI_15D

1 1 1

e9 e10 e11 e12

e5

e21 Model 2

e20

e18
1 1

e19
1 1 1

e4 e3 e2 e1

1 1 1

CY1 CY2 CY4 CY5

1

DE1 DE2 DE3 DE4 DE5

1 1 1 1

e13 e14 e15 e16 e17

MBI Cynicism
1

UWES Dedication
1

Distancing
1

Dedication

e8 1 e7
1

OLBI_3D OLBI_6D

1 1

OLBI_1D

e9

e6 OLBI_9D 1 1 e5 OLBI_11D

OLBI Disengagement
1

OLBI Dedication
1

OLBI_7D e10 1 OLBI_13D e11 1 OLBI_15D e12

e21 Model 3

e20

Figure 1. Hierarchical models of the structure of responses to all identication dimensions of burnout and work engagement.

216

DEMEROUTI, MOSTERT, AND BAKKER

Table 2 Goodness-of-Fit Indices (Maximum-Likelihood Estimates) for the Conrmatory Factor Analyses
Model 1. First-order factors 2. One second-order factor 3. Two second-order factors 4. Two second-order factors constrained Null 1. First-order factors 2. One second-order factor 3. Two second-order factors 4. Two second-order factors constrained Null
2

df

AGFI .90 .89 .90 .90 .26 .87 .87 .87 .87 .32

RMSEA .06 .07 .06 .06 .21 .07 .07 .07 .07 .20

TLI .92 .91 .92 .92 .87 .86 .87 .87

CFI .93 .92 .93 .93 .89 .88 .89 .88 root mean

Identication dimensions 332.90 113 .001 368.88 115 .001 334.01 114 .001 336.60 115 .001 3331.26 136 Energy dimensions 496.51 146 .001 521.18 148 .001 497.52 147 .001 507.35 148 .001 3242.47 171

Note. N 528. 2 chi square; df degrees of freedom; AGFI adjusted goodness of t index; RMSEA square error of approximation; TLI Tucker Lewis index; CFI comparative t index.

(Model 1). This would suggest that distancing and dedication are distinguishable (i.e., not representing two ends of a bipolar construct). However, the estimated correlation between the second-order factors was high (-.83). Indeed, the model (Model 4) that assumed no discriminant validity between the second-order factors, distancing and dedication, was not signicantly worse than Model 3, which included 2 (1) 2.59, two correlated second-order factors, ns, or Model 1, the rst-order factor model ( 2 (2) 3.70, ns). This suggests that distancing and dedication are opposite ends of one dimension supporting Hypothesis 1. The results for the energy dimension were somewhat different. Again Model 2, positing a single second-order factor showed a worse t to the data than Model 1. Model 3, positing two higher-order factors of exhaustion and vigor, did not t worse to 1.01, ns). This the data than Model 1 ( 2(1) indicates that exhaustion and vigor are distinguishable. The estimated correlation was high ( .81), which implies that exhaustion and vigor overlap substantially. Constraining the correlation between the higher order factors, exhaustion and vigor, to be equal to one resulted in a slightly worse t of the modelthat is, Model 4 was signicantly worse than both Model 3 ( 2(1) 9.83, p .01), including two distinguishable higher order factors, and Model 1, the rst-order factor model ( 2(2) 10.84, p .01). Thus, the energy components seem to form two distinguishable yet highly related dimensions.

Relations of Burnout and Work Engagement With Other Constructs

If burnout and work engagement are each others opposite, they should be equally strong related to other constructs but in the opposite direction. We focused on work pressure, autonomy, organizational commitment, and mental health. The examination is accomplished by adding each construct (as latent factors with manifest variables) separately to Model 2 (including one second-order factor) and Model 3 (including two second-order factors) considered in the previous section and by allowing them to correlate with the second-order factors. Additionally, age and gender were included as control variables with paths to each manifest variable. Table 3 displays the estimated standardized correlations. Work pressure. The work pressure latent factor was inferred from three item parcels (each representing the average of two items) as manifest variables. When work pressure was added to Model 2, it was unrelated to the identication second-order latent factor but was signicantly related to the second-order latent factor of vigor/exhaustion. When two secondorder latent factors were posited, work pressure was positively related to the exhaustion factor and unrelated to the vigor factor. Again, it was unrelated to the distancing and engagement factors. Thus, Hypothesis 3 is conrmed for work pressure since it is unrelated to both distancing and dedication. On the contrary, Hypothesis 4 should be rejected for work

BURNOUT AND WORK ENGAGEMENT

217

Table 3 Relations (Estimated Correlations) of Higher Order Energy and Identication Factors to Work Pressure, Autonomy, Organizational Commitment, and Mental Health After Controlling for Gender and Age
Models containing one higher order factor Energy1 Work pressure Autonomy Organizational commitment Mental health .20 .44 .48 .79 Identication1 .01 .41 .67 .56
#

Models containing two higher order factors Exhaustion .28 .38a .41a .77a
a

Vigor .02 .46b .55b .68b

#b

Distancing .10 .36a .59a .62a

#

Dedication .05# .40b .65b .47b

Note. All correlations were signicant at p .001 except for the correlations marked with the # symbol. 1 High scores indicate high work engagement (i.e. high energy and high identication level). a,b Means with different superscripts differ signicantly at the p .05 level (as calculated through AMOS by means of critical ratios for differences).

pressure because it shows differential relationships with the exhaustion and vigor factors. Autonomy. The autonomy latent factor was inferred from three item parcels (each representing the average of two items) as manifest variables. Autonomy was related to both the identication and the energy second-order latent factors. When two higher order factors of attitudes were posited (i.e., Model 2), the dedication-autonomy correlation was similar to the distancing-autonomy correlation. However, the vigor-autonomy correlation was signicantly higher than the exhaustion-autonomy correlation. Thus, autonomy showed the same pattern of relationships with both identication components, substantiating Hypothesis 3, but a more differentiated pattern for the two energy components, rejecting Hypothesis 4. Organizational commitment. Using the ve items we built three parcels to operationalize the latent factor of commitment. Organizational commitment had a stronger correlation with the identication factor than with the energy factor. When two higherorder identication factors were posited they showed a similar relationship with commitment. When two higher-order energy factors were included, the vigorcommitment correlation was signicantly stronger than the exhaustion-commitment correlation. Thus, similar to the ndings regarding autonomy, organizational commitment showed the same pattern of relationships with both identication components, substantiating Hypothesis 3, and a differentiated pattern for the two energy components, rejecting Hypothesis 4. Mental health. The mental health factor was inferred from four item parcels (each representing the average of seven items belonging to one dimension) as manifest variables. Mental health was signicantly

related to the second-order factors of attitudes and energy. However, the correlation was stronger for the energy factor. The model including separate exhaustion and vigor second-order factors showed that the correlation between exhaustion and mental health was stronger than the correlation between vigor and mental health. Similarly, mental health was stronger related to disengagement than to dedication. Contrary to Hypotheses 3 and 4, mental health is stronger related to the negatively worded dimensions.

Common Method Variance

As with all self-report data, there is the potential for the occurrence of method variance. Two tests were conducted to determine the extent of method variance in the current data. First, a Harmon onefactor test was conducted (Podsakoff & Organ, 1986) in two series of analysis: (a) all energy items and the items of each of the other constructs separately and (b) all identication items and the items of each of the other constructs. Results from these tests suggested the presence of at least ve factors in each analysis, indicating that common method effects were not a likely contaminant of the results observed in our study. To conrm these results, additional analyses were performed to test for common method variance following the procedure used by Williams, Cote, and Buckley (1989). We compared Model 3, including the additional constructs and the control variables, with a model including additionally a single method factor. Results indicated that while the method factor did improve model t in four of the seven cases (the model with energy items and work pressure could not be estimated), it accounted for a small portion (10%) of the total variance, which is

218

DEMEROUTI, MOSTERT, AND BAKKER

less than half the amount of method variance (25%) observed by Williams et al. (1989). Both tests suggest that common method variance is not a pervasive problem in this study.

Discussion
The aim of this study was to examine whether the dimensions of burnout and work engagement are bipolar constructs representing each others opposite. In order to investigate this we used the MBI-GS (measuring burnout using negatively formulated items only), the UWES (measuring work engagement using positively formulated items only), and the OLBI (measuring both burnout and work engagement as bipolar constructs using positively and negatively formulated items). Practically, these scales measure parallel dimensions using items with overlapping content. In addition, we examined the relationships of the derived dimensions to work pressure, autonomy, organizational commitment, and mental health. Taken together, the results inhibit us from providing a simple answer to the question whether burnout and work engagement are bipolar constructs. Our ndings indicate that we should answer this question for each dimension separately. While the identication dimensions of burnout (cynicism/disengagement) and work engagement (dedication) seem to be each others opposite, the energy dimensions (exhaustion vs. vigor) seem to represent two separate but highly related constructs. This conclusion can be justied both on the basis of the CFA ndings, and the pattern of relationships with other constructs. According to the CFA ndings, constraining the correlation between the second-order latent factors of the identication dimensions to be one did not make the model inferior to a model without this restriction. This means that their correlation was so high that we can assume that the constructs practically overlap. This nding agrees with Gonzalez-Roma et al.s (2006) ndings who used nonparametric methods to assess the dimensionality of two of the three instruments included in our study (MBI-GS and UWES). For the energy dimensions, however, constraining the correlation of the two second-order latent factors to one resulted in a signicantly worse model t. Although the bivariate and estimated correlation between exhaustion and vigor was high, they do not seem to form quite two opposites of one continuum. This nding also agrees with Gonzalez-Roma et al. (2006) who found that the exhaustion and vigor items constitute a weak to moderate energy dimension.

Findings regarding the relationships between the burnout and work engagement dimensions and hypothetical predictors and outcomes showed a similar picture. Expanding Gonzalez-Roma et al.s (2006) ndings, results showed that work pressure, autonomy, and organizational commitment have equally strong relationships with distancing and dedication, but in an opposite fashion. Only mental health turned out to be somewhat stronger related to distancing than to dedication. This does not seem to be an artifact of the item formulation because the GHQ-28 includes both positively and negatively worded items. These ndings largely support the idea that distancing and dedication represent a bipolar construct (identication) since they show no substantial differences in the pattern of relationships with other relevant constructs. In contrast, vigor and exhaustion show a different pattern of relationships with work pressure, autonomy, organizational commitment, and mental health. Autonomy and commitment are stronger related to vigor than to exhaustion, whereas work pressure and mental health are stronger related to exhaustion than to vigor. These ndings further substantiate the argument that vigor and exhaustion represent independent dimensions. The logical question now is how can we make sense of these ndings? The nding that the distancing and dedication factors represent two ends of one construct is not very surprising because people can either hold negative or positive attitudes toward their work. It seems unlikely that they endorse both simultaneously. This is also justied by the distribution of the scores across the identication dimensions. Thus, responses to the identication items of burnout and work engagement constructs seem to follow the structure of the circumplex of emotions as suggested by Watson and Tellegen (1985) where distancing and dedication are considered as two opposites of one continuum. In addition, Cacioppo and Berntson (1994) have argued that the evaluative space in which attitudes exist is two-dimensional, corresponding to the dimensions of the Watson and Tellegen model. On the contrary, the energy dimensions as operationalized by the various instruments seem to contain different aspects. This applies particularly to the operationalizations of vigor. While OLBI-vigor is measured with items like After working, I have enough energy for my leisure activities or When I work, I usually feel energized, a typical item of UWESvigor is At my work, I feel bursting with energy. The difference between these items is that OLBI conceives vigor as having sufcient energy reserves during and after work while UWES views vigor as

BURNOUT AND WORK ENGAGEMENT

219

having a surplus of energy reserves while being at work. Moreover, vigor, as dened by Schaufeli and Bakker (2003, 2004), in addition to the core meaning of high energy levels, seems to include a motivational element as well (i.e., the willingness to invest effort). Thus, conceptually and psychometrically, at least UWES-vigor is not exactly the opposite of exhaustion as measured with the MBI-GS because it also contains motivational aspects. In light of these ndings we could suggest that reporting different scores for the identication components of burnout and work engagement does not seem necessary since they more likely represent the same construct. Our ndings suggest using two different scores for MBI-exhaustion and UWES-vigor, because these scales measure two different but highly (negatively) related constructs. Alternatively, the OLBI instrument could be used, which has been proven to contain two factors of exhaustion and disengagement (or, positively framed, vigor and engagement) (Demerouti et al., 2003; Demerouti & Bakker, 2008) operationalized by positively and negatively worded items, thus capturing both ends of the continuum. Note that it is necessary to use the total scores for the exhaustion/vigor and for the engagement/ disengagement dimensions and not to split them as was done in the present study (cf. low reliability of OLBI vigor).

Limitations and Future Research

The rst limitation of the study is its reliance on self-report, cross-sectional data. While it provides a useful consideration of the factor structure of the different instruments, it cannot address the validity issues requiring a diversity of measurement formats. By conducting two different tests we found that responses to the items were not seriously inuenced by an articial common method factor. However, future studies aiming to examine dimensionality issues need to integrate data from other sources of information such as objective absenteeism in order to minimize common method artifacts. Several aspects of the study raise concerns regarding the generalizability of our results. Specically, although the sample of participants represented a diverse number of jobs (e.g., employees in different business units and departments), our sample is restricted to employees of the construction industry and has not been randomly selected from the full range of possible occupations. Moreover, our sample was overrepresented by White, middle-aged men. Future studies might focus more exclusively on other groups

in South Africa (e.g., Black, Colored, and Indian) from all age groups and in different sectors. However, the ndings seem generally consistent with Gonzalez-Roma et al. (2006) who conducted their research in The Netherlands with Dutch language instruments. Another possible drawback of this study is that the use of the English language for the questionnaires could also have a detrimental inuence on the results of the study because of the possibility of misunderstanding and misinterpretation of items from those participants for whom English is not their rst language. In order to minimize the inuence of this possible drawback, we explained the meaning of words that could have possible been misunderstood in footnotes. In order to reject the possibility that our ndings are inuenced by the instruments that we utilized, testing dimensionality issues with other scales would put our hypotheses to an even more robust test. However, the existing alternativesthat is, the instruments of Shirom (2003) and Kristensen et al. (2005) focus only on the exhaustion dimension. A related drawback concerns the low reliability ( .70) of the UWES-vigor scale and OLBI-vigor subscale. This might be due to the previous limitations, sampling error and misunderstanding of the items. Note, however, that the OLBI-vigor subscale is not supposed to be analyzed separately from OLBI-exhaustion. Together, the items form a reliable scale. A nal potential drawback concerns the way of analysis. First, as we conducted analysis for the energy and the identication dimensions separately, this has implications for establishing construct validity as for example, the relationships between the dimensions could not be controlled for. Second, when we conducted linearity tests of means comparison we found that of the 162 comparisons, 41 pairs showed a signicant deviation from linearity at p .003 (applying Bonferroni correction). In 21 of all signicant deviations from linearity, three items of UWES-vigor were involved. Strictly speaking, we would have to eliminate UWES-vigor from the analysis or conduct nonparametric analyses. However, because only one of the six scales seems to show nonlinear relationships with items of the other scales, we decided to keep this scale in the analysis and to continue with CFA instead of nonparametric tests. The practical importance of uncovering whether burnout and work engagement are each others opposite concerns mainly psychometric issues within organizational studies. Organizations need to have short and valid screening instruments to evaluate the

220

DEMEROUTI, MOSTERT, AND BAKKER loss spiral of work pressure, work-home interference and exhaustion: Reciprocal relations in a three-wave study. Journal of Vocational Behavior, 64, 131149. Demerouti, E., Bakker, A. B., Nachreiner, F., & Ebbinghaus, M. (2002). From mental strain to burnout. European Journal of Work and Organizational Psychology, 11, 423 441. Demerouti, E., Bakker, A. B., Nachreiner, F., & Schaufeli, W. B. (2001). The Job Demands-Resources Model of burnout. Journal of Applied Psychology, 86, 499 512. Demerouti, E., Bakker, A. B., Vardakou, I., & Kantas, A. (2003). The convergent validity of two burnout instruments: A multitrait-multimethod analysis. European Journal of Psychological Assessment, 19, 1223. Demerouti, E., & Nachreiner, F. (1996). Reliabilitat und Validitat des Maslach Burnout Inventory (MBI): Eine kritische Betrachtung [Reliability and validity of the Maslach Burnout Inventory: A critical note]. Zeitschrift fur Arbeitswissenschaft, 50, 3238. Demerouti, E., & Nachreiner, F. (1998). Zur Spezitat von Burnout fur Dienstleistungsberufe: Fakt oder Artefakt? [The specicity of burnout in human services: Fact or artifact?]. Zeitschrift fur Arbeitswissenschaft, 52, 82 89. Doty, D. H., & Glick, W. H. (1998). Common methods bias: Does common methods variance really bias results? Organizational Research Methods, 1, 374 406. Goldberg, R. J., & Williams, P. (1988). A users guide to the General Health Questionnaire. Windsor, U.K.: NFERNelson. Gonzalez-Roma, V., Schaufeli, W. B., Bakker, A. B., & Lloret, S. (2006). Burnout and work engagement: Independent factors or opposite poles? Journal of Vocational Behavior, 62, 165174. Hakanen, J. J., Bakker, A. B., & Schaufeli, W. B. (2006). Burnout and work engagement among teachers. Journal of School Psychology, 43, 495513. Halbesleben, J. R. B., & Demerouti, E. (2005). The construct validity of an alternative measure of burnout: Investigating the English translation of the Oldenburg Burnout Inventory. Work & Stress, 19, 208 220. Jackson, L. T. B., & Rothmann, S. (2005). Work-related well-being of educators in a district of the North-West Province. Perspectives in Education, 23, 107122. Karasek, R. A. (1985). Job content instrument: Questionnaire and users guide. Los Angeles: Dept. of Industrial and Systems Engineering, University of Southern California. Koekemoer, F. E., & Mostert, K. (2006). Job characteristics, burnout and negative work-home interference. South African Journal of Industrial Psychology, 32, 8797. Kristensen, T. S., Borritz, M., Villadsen, E., & Christensen, K. B. (2005). The Copenhagen Burnout Inventory: A new tool for the assessment of burnout. Work & Stress, 19, 192207. Lee, R. T., & Ashforth, B. E. (1990). On the meaning of Maslachs three dimensions of burnout. Journal of Applied Psychology, 75, 743747. Lee, R. T., & Ashforth, B. E. (1996). A meta-analytic examination of the correlates of the three dimensions of burnout. Journal of Applied Psychology, 81, 123133. Leiter, M. P., & Schaufeli, W. B. (1996). Consistency of the burnout construct across occupations. Anxiety, Stress and Coping, 9, 229 243. Le Roux, A. M. (2005). The validation of two burnout

occupational health of their employees. If burnout and work engagement can partly be conceived as each others opposites, this means that a fewer number of items are necessary to measure them. This implies that they have partly the same and partly different possible antecedents.

Conclusion
The present study offers evidence for the reliability and construct validity of a new instrument to assess burnout and work engagement. The Oldenburg Burnout Inventory (OLBI) captures the same constructs as assessed with the alternative measurement instruments MBI-GS (that assesses only burnout) and UWES (that assesses only work engagement). This means that the OLBI is a reasonable alternative that can be used to assess burnout and work engagement simultaneously. We hope that the present study encourages the use of the OLBI (see Appendix), but also further stimulates our understanding of the fascinating phenomena of burnout and work engagement.

References
Arbuckle, J. L. (2006). Amos (Version 7.0) [Computer Program]. Chicago: SPSS. Bagozzi, R. P. (1993). An examination of the psychometric properties of measures of negative affect in the PANAS-X scales. Journal of Personality and Social Psychology, 65, 836 851. Bakker, A. B., Demerouti, E., & Schaufeli, W. B. (2002). The validity of the Maslach Burnout Inventory - general survey: An internet study. Anxiety, Stress, and Coping, 15, 245260. Bakker, A. B., Demerouti, E., & Verbeke, W. (2004). Using the job demands-resources model to predict burnout and performance. Human Resource Management, 43, 83 104. Bentler, P. M., & Chou, C. P. (1987). Practical issues in structural modeling. Social Methods & Research, 16, 78 117. Cacioppo, J. T., & Berntson, G. G. (1994). Relationship between attitudes and evaluative space: A critical review, with emphasis on the separability of positive and negative substrates. Psychological Bulletin, 115, 401 423. Demerouti, E. (1999). Burnout: Eine Folge Konkreter Abeitsbedingungen bei Dienstleistungs und Produktionstdtigkeiten. (Burnout: A consequence of specic working conditions among human service and production tasks). Frankfurt/Main: Lang. Demerouti, E., & Bakker, A. B. (2008). The Oldenburg Burnout Inventory: A good alternative to measure burnout and engagement. In J. Halbesleben (Ed.), Stress and burnout in health care (pp. 6578). Hauppage, NY: Nova Sciences. Demerouti, E., Bakker, A. B., & Bulters, A. J. (2004). The

BURNOUT AND WORK ENGAGEMENT measures in the South African earthmoving equipment industry. Unpublished masters dissertation, North-West University, Potchefstroom Campus, South Africa. Lewig, K. A., Xanthopoulou, D., Bakker, A. B., Dollard, M. F., & Metzer, J. C. (2007). Burnout and connectedness among Australian volunteers: A test of the job demands-resources model. Journal of Vocational Behavior, 71, 429 445. Llorens, S., Bakker, A. B., Schaufeli, W., & Salonova, M. (2006). Testing the robustness of the Job Demands Resources model. International Journal of Stress Management, 13, 378 391. Marsh, H. W., Antill, J. K., & Cunningham, J. D. (1989). Masculinity and femininity: A bipolar construct and independent constructs. Journal of Personality, 57, 625 663. Maslach, C., Jackson, S. E., & Leiter, M. P. (1996). Maslach Burnout Inventory manual (3rd ed.). Palo Alto, CA: Consulting Psychologists Press. Maslach, C., & Leiter, M. P. (1997). The truth about burnout. San Francisco: Jossey-Bass. Maslach, C., Schaufeli, W. B., & Leiter, M. P. (2001). Job burnout. Annual Review of Psychology, 52, 397 422. Mauno, S., Kinnunen, U., & Ruokolainen, M. (2007). Job demands and resources as antecedents of work engagement: A longitudinal study. Journal of Vocational Behavior, 70, 149 171. Meyer, J. P., Allen, N. J., & Smith, C. A. (1993). Commitment to organizations and occupations: Extension and test of a three-component model. Journal of Applied Psychology, 78, 538 551. Podsakoff, P. M., & Organ, D. W. (1986). Self-reports in organizational research: Problems and prospects. Journal of Management, 12, 531544. Price, J. L., & Mueller, C. W. (1986). Handbook of organizational measurement. Marsheld, MA: Pitman. Richardsen, A. M., & Martinussen, M. (2004). The Maslach Burnout Inventory: Factorial validity and consistency across occupational groups in Norway. Journal of Occupational and Organizational Psychology, 77, 120. Rost, I. (2007). Work wellness of employees in the earthmoving equipment industry. Unpublished doctoral thesis, North-West University, Potchefstroom Campus, South Africa. Rothmann, S., & Pieterse, J. (2007). Predictors of workrelated well-being in sector education training authorities. South African Journal of Economic and Management Sciences, 10, 298 312. Schaufeli, W. B., & Bakker, A. B. (2003). UWES - Utrecht Work Engagement Scale: Test Manual. Utrecht, The Netherlands: Department of Psychology, Utrecht University. Schaufeli, W. B., & Bakker, A. B. (2004). Job demands, job resources and their relationship with burnout and engage-

221

ment: A multi-sample study. Journal of Organizational Behavior, 25, 293315. Schaufeli, W. B., & Bakker, A. B. (2010). Dening and measuring work engagement: Bringing clarity to the concept. In A. B. Bakker & M. P. Leiter (Eds.), Work engagement: A handbook of essential theory and research (pp. 10 24). New York: Psychology Press. Schaufeli, W. B., Leiter, M. P., Maslach, C., & Jackson, S. E. (1996). The Maslach Burnout Inventory-General Survey. In C. Maslach, S. E. Jackson, & M. P. Leiter (Eds.), Maslach Burnout Inventory. Palo Alto, CA: Consulting Psychologists Press. Schaufeli, W. B., Salanova, M., Gonzalez-Roma, V., & Bakker, A. B. (2002). The measurement of engagement and burnout: A two sample conrmatory factor analytic approach. Journal of Happiness Studies, 3, 7192. Schutte, N., Toppinen, S., Kalimo, R., & Schaufeli, W. B. (2000). The factorial validity of the Maslach Burnout InventoryGeneral Survey (MBIGS) across occupational groups and nations. Journal of Occupational and Organizational Psychology, 73, 53 66. Shirom, A. (2003). Job-related burnout. In J. C. Quick, & L. E. Tetrick (Eds.), Handbook of occupational health psychology (pp. 245/265). Washington, DC: American Psychological Association. Sonnentag, S. (2003). Recovery, work engagement, and proactive behavior: A new look at the interface between non-work and work. Journal of Applied Psychology, 88, 518 528. Storm, K., & Rothmann, S. (2003). A psychometric analysis of the Maslach Burnout Inventory-General Survey in the South Africa Police Service. South African Journal of Psychology, 33, 219 226. Van Abswoude, A. A. H., Vermunt, J. K., Hemker, B. T., & Van der Ark, L. A. (2004). Mokken scale analysis using hierarchical clustering procedures. Applied Psychological Measurement, 28, 332354. Van Veldhoven, M., Meijman, T. F., Broersen, J. P. J., & Fortuin, R. J. (1997). Handleiding VBBA: Onderzoek naar de beleving van psychosociale arbeidsbelasting en werkstress met behulp van de vragenlijst beleving en beoordeling van de arbeid [Manual VBBA: Research on the experience of psychosocial work load and job stress with the Questionnaire on the Experience and Evaluation of Work]. Amsterdam: SKB. Watson, D., & Tellegin, A. (1985). Toward a consensual structure of mood. Psychological Bulletin, 98, 219 235. Williams, L. J., Cote, J. A., & Buckley, M. R. (1989). Lack of method variance in self-reported affect and perceptions at work: Reality or artifact? Journal of Applied Psychology, 74, 462 468. Yi-Wen, Z., & Yi-Qun, C. (2005). The Chinese version of Utrecht Work Engagement Scale: An examination of reliability and validity. Chinese Journal of Clinical Psychology, 13, 268 270.

(Appendix follows)

222

DEMEROUTI, MOSTERT, AND BAKKER

Appendix Oldenburg Burnout Inventory

Instruction: Below you nd a series of statements with which you may agree or disagree. Using the scale, please indicate the degree of your agreement by selecting the number that corresponds with each statement
Strongly Strongly agree Agree Disagree disagree 1. I always nd new and interesting aspects in my work. 2. There are days when I feel tired before I arrive at work. 3. It happens more and more often that I talk about my work in a negative way. 4. After work, I tend to need more time than in the past in order to relax and feel better. 5. I can tolerate the pressure of my work very well. 6. Lately, I tend to think less at work and do my job almost mechanically. 7. I nd my work to be a positive challenge. 8. During my work, I often feel emotionally drained. 9. Over time, one can become disconnected from this type of work. 10. After working, I have enough energy for my leisure activities. 11. Sometimes I feel sickened by my work tasks. 12. After my work, I usually feel worn out and weary. 13. This is the only type of work that I can imagine myself doing. 14. Usually, I can manage the amount of my work well. 15. I feel more and more engaged in my work. 16. When I work, I usually feel energized. 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4

Note. Disengagement items are 1, 3(R), 6(R), 7, 9(R), 11(R), 13, 15. Exhaustion items are 2(R), 4(R), 5, 8(R), 10, 12(R), 14, 16. (R) means reversed item when the scores should be such that higher scores indicate more burnout.

Received July 15, 2008 Revision received October 3, 2009 Accepted October 12, 2009 y

E-Mail Notication of Your Latest Issue Online!

Would you like to know when the next issue of your favorite APA journal will be available online? This service is now available to you. Sign up at http://notify.apa.org/ and you will be notied by e-mail when issues of interest to you become available!