The Relationship of Burnout to Work Group Psychological Safety Among Public Child Welfare Workers: A Multilevel Perspective

Background: The significant cost of burnout to child welfare workers, clients, and agencies has led to a large literature aimed at identifying its antecedents. However, virtually all burnout studies have restricted themselves to using only individual level variables, thus eliminating the possibility of identifying how processes at various levels of the organization potentially affect the individual. This study includes social support measures frequently found to be significantly related to burnout levels and introduces a measure of work group psychological safety, which reflects a shared evaluation that the work unit is a supportive environment in which it is safe to raise difficult issues and take risks.   Results indicate team psychological safety has a dynamic relationship with burnout; in some units the work unit matters more than others. The results of this analysis highlight the complex nature of work group interventions and point toward the value of considering targeted work group interventions for addressing burnout.

Methods: Direct-service social workers employed by a child welfare agency in the Pacific Northwest were surveyed using a combination of paper-based forms and web-based surveys. The overall response rate was 96%. The current analysis focused on a subset of social workers, case-carrying workers who had regular, direct, face-to-face contact with families (n=1040).  The measurement of psychological safety used the original seven item scale developed by Edmondson (1999). Burnout was measured using a variation of the depersonalization and emotional exhaustion subscales of burnout developed by Maslach (1981), with all three measures utilizing a five point Likert scale. The polychoric coefficient alpha was .86 for psychological safety, .90 for depersonalization, and .83 for emotional exhaustion. Our initial descriptive analysis revealed that a significant proportion of the variance in team psychological safety was accounted for at the level of the supervisory unit. This led us to test a multilevel structural equation model in which we examined the individual and unit-level effects of team psychological safety on our measures of burnout.

Results: The initial results of our multilevel structural equation model reveal significant and negative effects of psychological safety on worker burnout. Some of this effect results directly from a worker’s individual perceptions of their work unit. In some units, however, the work unit significantly affects a worker's level of psychological safety and this “unit-level effect” appears to be carried forward into a worker’s level of burnout. In this way, the manner in which team psychological safety impacts a worker’s level of burnout can be said to “depend” on the worker’s unit. The multilevel dynamics of this relationship will be highlighted by using a graphical analysis in the presentation.

Conclusion: Studies of predictors of burnout have almost exclusively been based on only individual-level factors. These results support the importance of work group level contextual factors to gain a more complete understanding of the antecedents to burnout.  The results of the analysis also suggest the value of considering both work group interventions as well as at the individual level as means of improving employee and organizational outcomes.