Freudenberger (Reference Freudenberger1974) coined the term burnout as a state of mental and physical exhaustion caused by worker’s professional life. Thus, individuals who burn out from their work would deplete their energetic resources and lose their dedication to work (Bakker et al., Reference Bakker, Demerouti and Sanz-Vergel2014). Burnout was initially thought to describe human service workers’ feeling of emotional exhaustion, depersonalization, and reduced personal accomplishment (Taris, Schreurs, & Schaufeli, Reference Taris, Schreurs and Schaufeli1999). However, the concept has now expanded to all workers who interact with people to some extent. Accordingly, burnout may be defined as a syndrome characterized by emotional exhaustion, depersonalization and lack of personal accomplishment (Bakker et al., Reference Bakker, Demerouti and Sanz-Vergel2014).

On the other hand, work engagement may be seen as a natural criterion for the validity of burnout factors, as work engagement is defined as a positive, accomplished, and encouraged attitude towards the job. According to this, Maslach and Leiter (Reference Maslach and Leiter1997) stated that work engagement may be conceptualized as the opposite pole of burnout. Indeed, there are two theoretical positions on the nature of both constructs. One side argues that both burnout and work engagement are two opposite poles of the same continuum, while others consider the construct to be an independent concept, that shows a negative relationship with burnout. On one hand, Maslach and Leiter (Reference Maslach and Leiter1997) considered burnout and work engagement as two parts of the same composition. According to these authors, burnout would be composed by emotional exhaustion, cynicism, and reduced professional efficacy, while work engagement would be defined by vigor, involvement and efficacy. On the other hand, Schaufeli and Bakker (Reference Schaufeli and Bakker2001) defended that burnout and work engagement constituted two different concepts.

Emotional exhaustion relates to being emotionally depleted of your emotional resources (Taris et al., Reference Taris, Schreurs and Schaufeli1999). Depersonalization or cynicism refers to “a negative or excessively detached response toward these people, who are the recipients of one’s service or care” (Bakker et al., Reference Bakker, Demerouti and Sanz-Vergel2014, p. 390). Finally, the third dimension referred to a reduced personal accomplishment, can be defined as the “tendency to evaluate one’s work negatively. Such an evaluation is often accompanied by feelings of insufficiency, self-doubt and poor self-esteem” (Xanthopoulou, Bakker, Demerouti, & Kantas, Reference Xanthopoulou, Bakker, Demerouti and Kantas2012, p. 41).

Several measures or instruments for the measurement of burnout have been developed. With no intention to be exhaustive, the Oldenburg Burnout Inventory (OLBI; Demerouti & Nachreiner, Reference Demerouti and Nachreiner1998), the Staff Burnout Scale for Health Professionals (Jones, Reference Jones1980), the Burnout Measure (Malakh-Pines & Aronson, Reference Malakh-Pines and Aronson1988), the Shirom-Melamed Burnout Measure (SMBM; Shirom & Melamed, Reference Shirom and Melamed2006), and the Copenhagen Burnout Inventory (Kristensen, Borritz, Villadsen, & Christensen, Reference Kristensen, Borritz, Villadsen and Christensen2005). Among them, the OLBI covers both affective aspects of exhaustion, and also physical and cognitive aspects (Demerouti, Bakker, Vardakou, & Kantas, Reference Demerouti, Bakker, Vardakou and Kantas2003). Analyses on German employees revealed a two correlated factor structure: exhaustion and disengagement. Another recently developed scale is the Copenhagen Burnout Inventory, which focuses on exhaustion as the most important element of burnout. This instrument is integrated by three scales: personal burnout, work-related burnout, and client-related burnout (Borritz et al., Reference Borritz, Rugulies, Bjorner, Villadsen, Mikkelsen and Kristensen2006). There are also some measures originally developed for Spanish-speaking populations. For example, the Spanish Burnout Inventory is composed by four sub-scales, namely enthusiasm toward the job, psychological exhaustion, indolence, and feeling guilty (Gil-Monte & Zúñiga-Caballero, Reference Gil-Monte and Zúñiga-Caballero2009).

However, the most widely used scale for measuring the components of burnout is, by far, the Maslach Burnout Inventory in its different versions: MBI-Human Services Survey, MBI-Educators Survey, MBI-General Survey, and MBI-Student Survey (Maslach, Jackson, & Leiter, Reference Maslach, Jackson and Leiter1996; Schaufeli, Martinez, Pinto, Salanova, & Bakker, Reference Schaufeli, Martinez, Pinto, Salanova and Bakker2002). Among the available versions of the scale, the Maslach Burnout Inventory-General Survey (MBI-GS; Schaufeli, Leiter, Maslach, & Jackson, Reference Schaufeli, Leiter, Maslach, Jackson, Maslach, Jackson and Leiter1996) was developed to measure three dimensions of burnout (cynicism, personal accomplishment, and emotional exhaustion) regardless of the type of work. It is composed of 16 items, five of them corresponding to cynicism, five others resembling emotional exhaustion, and six assessing personal accomplishment. The MBI-GS differs from the original MBI in that the former includes a more general view about the aspects of the job, which is not restricted to their social features; nevertheless, the sub-scales resemble those of the original form.

Although being an extremely popular measure of burnout, the MBI in its different versions is not free of criticism for a number of reasons. For the particular case of the MBI-GS, one of them is that the MBI-GS presents the same psychometric limitations as its original version “namely, that the items in each subscale are all framed in the same direction” (Demerouti et al., Reference Demerouti, Bakker, Vardakou and Kantas2003, p. 297). This implies that the “negative” factors of the scale (cynicism and emotional exhaustion) are composed by negatively worded items, and by contrast, the positive factor, or personal accomplishment, is positively worded. Also, the Maslach Burnout Inventory has raised some doubts on its validity, as there has been disagreement regarding the number of factors that make up the construct. Some authors have suggested that the factor structure for the MBI is not stable across occupational groups (Leiter & Schaufeli, Reference Leiter and Schaufeli1996). In particular, the depersonalization and emotional exhaustion subscales tended to collapse into one factor when groups other than human service providers completed the MBI (Bakker, Demerouti, & Schaufeli, Reference Bakker, Demerouti and Schaufeli2002). That was indeed one reason to develop the MBI-GS. The MBI-GS assumes the three-factorial hypothesis by which emotional exhaustion and cynicism correlate and personal accomplishment is shown as an independent factor. Additionally, there are two other factorial models that try to explain the concept of burnout and which can be found along literature. The first of them hypothesizes a unidimensional structure of burnout (Malakh-Pines & Aronson, Reference Malakh-Pines and Aronson1988). Green, Walkey, and Taylor (Reference Green, Walkey and Taylor1991) found a two-factor structure with one dimension formed by emotional exhaustion and cynicism, and the other conformed by personal accomplishment.

The MBI-GS is a relatively young instrument and thus there is limited literature about its validations. However, it is worth noting that the measure has been validated in Canada (Leiter & Schaufeli, Reference Leiter and Schaufeli1996), in Sweden (Schutte, Toppinen, Kalimo, & Schaufeli, Reference Schutte, Toppinen, Kalimo and Schaufeli2000), in the Netherlands (Bakker et al., Reference Bakker, Demerouti and Schaufeli2002), in Japan (Kitaoka-Higashiguchi, Ogino, & Masuda, Reference Kitaoka-Higashiguchi, Ogino and Masuda2004), in Finland (Ahola et al., Reference Ahola, Honkonen, Isometsä, Kalimo, Nykyri, Koskinen and Lönnqvist2006), in Norway (Langballe, Falkum, Innstrand, & Aasland, Reference Langballe, Falkum, Innstrand and Aasland2006), and in Poland (Chirkowska-Smolak & Kleka, Reference Chirkowska-Smolak and Kleka2011).

With respect to the Spanish version of the scale, the instrument has been validated in Spain (Gil-Monte, Reference Gil-Monte2002; Moreno-Jiménez, Rodríguez-Carvajal, & Escobar, Reference Moreno-Jiménez, Rodríguez-Carvajal and Escobar Redonda2001), in Cuba (Oramas, González, & Vergara, Reference Oramas, González and Vergara2007), and Venezuela (Millan & d’Aubeterre, Reference Millan and d’Aubeterre2012). These four validations studied the factor structure by exploratory means, most of them with orthogonal rotations (i.e., uncorrelated factors) and all of them supported the theoretical three-factor solution. A confirmatory study on the Spanish version of the MBI-GS by Salanova, Schaufeli, Llorens, Peiró, and Grau (Reference Salanova, Schaufeli, Llorens, Peiró and Grau2000) found some evidence for a three factor structure. Nevertheless, this research presented some shortcomings: an incidental sample; the use of a estimation method (Maximum Likelihood) that assumes multivariate normality and continuous variables in a questionnaire with ordinal variables usually non-normal; and more important, the original three factor structure did not fit well, and in order to improve model fit four error covariances were introduced. Covarying covariances has been extensively criticized as having very little theoretical basis and inflating the fit (for example, Kline, Reference Kline2011).

All in all, the MBI has been subjected to extensive research and has a well-documented history of validity and reliability studies. Nevertheless, the MBI-GS has not been so widely validated, as seen in the aforementioned literature. This is even clearer with the Spanish version of the MBI-GS. First, the validation studies have been exploratory, with no theoretical reasons to use principal component analysis and orthogonal rotations, instead of correlated factors as defended by the authors of the scale. The available confirmatory results in a Spanish sample (Salanova et al., Reference Salanova, Schaufeli, Llorens, Peiró and Grau2000) did not fit the data well until four error covariances were included. All this makes it extremely difficult to compare the relative fit of the different structures proposed in the literature. Second, most of the validations have used non-probabilistic samples. Additionally, and to the knowledge of these researchers, the MBI-GS has never been validated in the Dominican Republic. Therefore, the aim of this research is to offer evidence on the psychometric properties of the MBI-GS, specifically reliability, factorial and criterion-related validities, for its use in the Dominican Republic and potentially other Caribbean Spanish-speaking countries, using a large and representative sample of Dominican teachers, as well as offering normative data to be used with these professionals.

Results

Factorial Validity

The four models shown in the Figure 1 were examined in order to test their validity. The first three models are completely a priori, and correspond to: Model 1, a unidimensional measure of burnout; Model 2, a two-factor structure of burnout (professional efficacy and the core of burnout); and Model 3, the theoretical three-factor structure (professional efficacy, cynicism, and emotional exhaustion). Finally, a fourth model was tested which is a slight ad hoc modification of model 3 that removes item 11 (an indicator of cynicism) given its low factor loading on the intended factor.

Figure 1.

Alternative confirmatory factor models

Table 1 shows the fit indices for the aforementioned four models. Best fitting model among the a priori ones was model 3. Model 3 corresponds to the theoretical structure of the MBI-GS. Nevertheless, item 11 had an extremely low factor loading on the cynicism dimension (λ = –.13) and was therefore removed. That yield to model 4, which was in turn the best fitting model. Standardized and unstandardized factor loadings for model 4, the one retained, are presented in Table 2.

Table 1.

Goodness-of-fit indices for the tested models

Note: **p < .01.

Table 2.

Unstandardized and standardized factor loadings for the best-fitting model

Note: **p < .01.

Internal consistency

Cronbach’s alphas for the three-factor structure found via CFA for the Spanish version of the MBI-GS were .67 for emotional exhaustion, .77 for professional efficacy and .48 for cynicism. When item 20 was removed from the cynicism dimension, its alpha estimate increased up to .54. As well as alpha, CRI’s were also calculated, and results were .78 for emotional exhaustion, .86 for professional efficacy, and .65 for cynicism. When item 11 was removed from the cynicism dimension, its internal consistency increased up to .74. According to the CRI, internal consistency seems adequate. Once dimensionality and internal consistencies of those dimensions is shown to be adequate, norms can be established, given that the sample is representative of the overall population of teachers in the Dominican Republic. These norms are presented in Annex I.

Criterion-related validity

Relations among job burnout and work engagement provide evidence of criterion-related validity, as both constructs have been hypothesized as opposite poles or, at least, as highly related. Table 3 shows the correlations among the dimensions of burnout (emotional exhaustion cynicism, and professional efficacy) and the three dimensions of work engagement (vigor, dedication, and absorption). As expected, the correlations among the dimensions of burnout and those of work engagement were statistically significant and in the expected direction.

Table 3.

Correlations among the dimensions of job burnout and work engagement

* p < .05; **p < .01.