Worker Characteristics and Their Associations with Family Group Conference Referrals: An Exploration into Implementation Variations

Background and Purpose. While policies are intended to guide behavior and decisions, staff are often left some degree of discretion with respect to their implementation. Results from a recent randomized control trial examining the efficacy of family group conferences in preventing the placement of children in out-of-home care, identified notable inconsistencies in the degree to which staff followed a protocol for referring a family to a family group conference (FGC). In order to understand why implementation of the protocol varied between staff, results from a staff survey were linked to the number of FGC referrals they completed during the study.

Methods: A web-based GSS was administered to and received from all frontline staff working or newly hired in the study sites throughout the course of the study. The survey included questions concerning tenure and duties, skills, job satisfaction, family group decision making knowledge and attitudes, perceptions of organizational culture and climate, services, and demographic information. Results for Poisson regression models were derived using generalized estimating equations with an exchangeable correlation structure to account for the clustering of data within units; the number of FGC referrals was the dependent variable. Survey data from 122 workers and data reflecting the number of family referrals to FGCs (obtained from a log that tracked randomization and referrals) were analyzed.

Results: The results indicate that worker referrals to FGCs varied significantly both between workers and work units, despite policies around referral guidelines. Further, time in current position and perceptions of supervisor competence and leadership effectiveness were significantly associated with a worker’s propensity to refer to FGC. Specifically, caseworkers with fewer years in their current position and those who reported lower ratings on their supervisor’s competence and leadership’s effectiveness made a higher number of referrals. Gender, age, and study site did not appear to influence FGC referrals.

Conclusions and Implications: Caseworkers, like other street-level bureaucrats, often have some degree of discretion in deciding what services families receive, despite the existence of policies or procedures that may be in place to guide or constrain their decisions. The findings from this study suggest that personal characteristics may be associated with a threshold for FGC referral and that the FGC may be a mechanism by which newer workers and/or those feeling less supervisory and leadership support meet a need that is not being adequately addressed by their supervisors and/or agency leadership. The results reinforce the notion that even in the presence of policy mandates, such factors should be monitored and addressed by agencies concerned with implementation fidelity and evaluative rigor.