Variables associated with turnover in 110 small Minnesota group homes supporting people with developmental disabilities between December 1993 and December 1996 were examined. Common staffing problems included recruiting qualified workers (reported by 57% of supervisors) and staff turnover (44%). Turnover for direct support professionals was 46%. Of workers who left, 45% left within 6 months of hire, and another 23%, between 6 and 12 months. A multiple regression analysis (with 11 variables) accounted for 34% of the variability in facility turnover rates. Variables contributing significantly to the prediction of higher turnover included greater support needs of residents, lower starting pay, less tenured supervisors, and fewer direct support professionals eligible for paid leave. Implications for practice, research, and policy were discussed.
The preparation of this manuscript was supported by Subcontracts No. H133B0003-90 and H133D50037-96 between the National Institute on Disability and Rehabilitation Research (NIDRR); the Research and Training Center on Residential Services and Community Living, Institute on Community Integration (UAP), University of Minnesota; and the Research and Training Center on Community Integration, Center on Human Policy at Syracuse University. Additional support was provided by NIDRR Grants No. H133B80048 and H133B0072-93 to the RRTC at the University of Minnesota. This manuscript is based on a dissertation submitted by the first author, based on a research project supervised by the second author, to the Department of Educational Psychology, University of Minnesota. The authors acknowledge the dissertation committee for their feedback and support: Robert H. Bruininks (chair), Susan Rose, Richard Arvey, Jennifer York, Richard Weatherman, and James Brown. They also acknowledge the contributions of Amy Hewitt, Janet Bast, Stacey Tytler-Moore, Erin Simond, and Lynda Anderson, who assisted with the literature review and data collection, and Nahoon Kwak, who developed system files for this project and provided general statistical support. A comprehensive report on the broader study was published as a monograph (Larson, Lakin, & Bruininks, 1998).