Discussion 6 1 Work engagement versus burnout
Discussion 6 1 Work engagement versus burnout
Burnout (BU) and work engagement (WE) have significant implications for employee health and organizational performance (e.g., Taris, 2006; Christian, Garza, & Slaughter, 2011), therefore they represent high-interest topics for researchers and practitioners. However, because the two concepts are rather highly correlated (Halbesleben, 2010), the relationship between BU and WE has generated debates in the literature. Initially, researchers considered that WE is the opposite of BU (Maslach & Leiter, 1997; Cole, Walter, Bedeian, & O’Boyle, 2012), and that both concepts can be assessed using the same questionnaire. In response to this perspective, other research studies showed that BU and WE have different correlation patterns with variables of interest (e.g., job characteristics) (Schaufeli, Taris, & van Rhenen, 2008); that WE has incremental effects over BU in longitudinal studies (e.g., Hakanen & Schaufeli, 2012); or that BU and WE have different correlation patterns with personality variables such as neuroticism or extraversion (Langelaan, Bakker, Van Doornen, & Schaufeli, 2006). Based on these findings, researchers concluded that BU and WE are constructs that describe connected, yet distinct forms of well-being (Schaufeli & Salanova, 2014).
In the present review, we start from the assumption that BU and WE are distinct and yet correlated forms of well-being. Following this conceptualization, some researchers suggested that the strong correlation between them (i.e., values ranging between 0.30 and 0.50, according to Halbesleben, 2010) could be the result of a causal relationship between the two forms of well-being. For example, Van Beek, Kranenburg, Taris, and Schaufeli (2013) suggested that highly engaged students are less vulnerable to exhaustion (a BU component), as compared with students with low engagement. Consequently, Van Beek et al. (2013) considered that WE is an antecedent of low exhaustion. Nonetheless, based on longitudinal designs, other researchers reported that rather BU is a significant predictor of (low) WE (Salmela-Aro & Upadyaya, 2014). To the best of our knowledge, these divergent perspectives were not previously addressed in a systematic manner. Moreover, most studies addressed the relationship between the two concepts based on a cross-sectional methodology, which makes it impossible to investigate causal relationships. Except for Salmela-Aro and Upadyaya (2014), longitudinal research studies focus their analyses on understanding causal relationships between well-being (BU and WE) and various outcomes (e.g., performance), and not on the reciprocal relationships between BU and WE. Therefore, we aim to clarify the relationship between BU and WE, using the data reported by longitudinal studies. To achieve this goal, we integrated meta-analytical calculations to combine results from different longitudinal studies and structural equation modeling procedures to test different cross-lagged models of the possible temporal order of BU and WE.
Burnout is characterized by three dimensions (Maslach & Leiter, 1997), namely exhaustion, cynicism (or depersonalization), and inefficacy (or reduced personal accomplishment). Specifically, exhaustion refers to the feeling of being drained, emotionally and physically, having low levels of energy; cynicism is conceptualized as a detached attitude towards work or people at work. Importantly, Bresó, Salanova, and Schaufeli (2007) draw attention to the problem of the third dimension of BU, initially called personal accomplishment, due to its positively worded items, and proposed an alternative dimension, named professional inefficacy, with negatively worded items, that should be used to measure burnout. Previous meta-analyses (e.g., Lee & Ashforth, 1996) indicated that inefficacy (measured as personal accomplishment) is relatively weakly correlated with exhaustion and cynicism (correlation values around 0.35, after reliability corrections) Consequently, researchers suggested that exhaustion and cynicism constitute a general factor, called core burnout (Green, Walkey, & Taylor, 1991).
To summarize, burnout is a multi-dimensional construct consisting of three dimensions that are not very strongly correlated. Therefore, in the present review, we will compute separate correlation values for the relations of each BU dimension and WE.
