[A meta-analysis of job burnout using the MBI scale].

BACKGROUND AND AIM: The aim of this study was to investigate the theoretical dimensionality of burnout measurement (MBI). The objective of the study was to check the working hypothesis according to which the subscales of the Maslach Burnout Inventory (MBI) such as "emotional exhaustion" and "cynicism" (depersonalization) were not correlated with that of "personal achievement" (professional efficacy). To do so, we performed a meta-analysis [Maslach and Jackson, The Maslach Burnout Inventory, 1981, 1986].

METHODS: The sample included 12,112 participants (sample size range from 100-3312 participants; estimated mean=448.59; standard deviation=648.51). Doctheses, Francis, Interscience, Kluweronline, Medline, PsycInfo, and ScienceDirect were searched to identify papers. Exploring references identified 83 publications (1998-2001). Two keywords were used: "Maslach Burnout Inventory" and "organizational stress". The criteria for this meta-analysis included: using the MBI "human services survey (HSS)" for professional care givers, "educators' survey (ES)" for professional educators, "general survey (GS)" for other workers).

RESULTS: About 58% of the population concerned professional caregivers, 13% professional educators, 11% social workers, and 18% correctional officers and other workers. In this meta-analysis, effect sizes should be reported with the number of studies and confidence intervals to test the consistency and reliability of the mean estimated (effect size). In this case, each correlation can be corrected. The meta-analysis can therefore be conducted on this corrected correlation. The aim will be to control any artefact and sampling error. Meta-analysis showed that the value of the corrected correlation is not always negative in the case of a relation between the dimension of "emotional exhaustion" and "personal achievement" (professional efficacy). In the case of the dimension "emotional exhaustion" and "cynicism" (depersonalization), we observed that the value of the corrected correlation studies and confidence intervals showed that this dimension was always correlated positively in the studies included. For the relationship between the value of corrected correlation studies and confidence intervals of the dimension of "cynicism" (depersonalization) and "personal achievement" (professional efficacy), the studies included showed that they were always correlated negatively. These results confirm the hypothesis studied.

CONCLUSION: The study revealed two points: (1) the homogeneity of the studies included concerning the fact that assessment of emotional exhaustion and cynicism (depersonalization) are always correlated positively; (2) the heterogeneity of the studies included concerning the fact that personal achievement is correlated with the other subscales of burnout (MBI). This study confirms the hypothesis.