4. Discussion
Because of the sustained debate regarding the distinctiveness of burnout and depression and due to the lack of a conclusive definition of this construct, the aim of the present study was to test whether life stress accounts for the relationship between depression and burnout. As there is the widespread assumption that depression arises from and affects many areas of life, whereas burnout is – not only but mainly – restricted to job stressors (Hakanen & Schaufeli, 2012), we assessed experienced life stress in a large sample of employed adults and in a sample of depressed inpatients who were not employed during participation. We hypothesized that both constructs are related, but that the number of stressful life events is closer related to depression than to burnout. In our data, we found a moderate correlation between burnout and depression, which is quite similar to that reported by Toker et al. (2005) who found correlation coefficients of about r = .50 between burnout and depression for men and for women. With an amount of about 36% of shared variance in the healthy control sample, we found a closer relation than other authors suggested in the past (e.g., Iacovides et al., 2003). However, the association was still moderate which means that important shares of variance were not equal for both constructs. Furthermore, correlations of the three MBI subscales drew the typical picture of closest relation between depression and emotional exhaustion and lowest association between depression and reduced efficacy. These findings are in line with shared variance estimates by Schaufeli and Enzmann (1998), who found an overlap with depression of 12–38% (emotional exhaustion), 2–29% (depersonalization), and 3–20% (reduced efficacy) respectively. In the clinical sample, there was an amount of shared variance of about 20%, which meets the shared variance estimates of the literature very well. Worthy of note, Bianchi et al. (2014) also reported similar correlations between depression and burnout for their whole sample but report a much closer relation when only taking highly burned-out participants into account. They argue that the fact that most studies only include healthy participants might lead to an underestimation of the correlation between burnout and depression as these results cannot be generalized to a clinical setting. However, our data suggest the same underlying mechanisms in the healthy as well as in the inpatient depressed sample questioning this assumption. Moreover, the correlation between both constructs was smaller in the depressed sample than in the healthy sample (r = .456 vs. r = .613).