Risk was measured with the Washington Risk Assessment Matrix using canonical correlation techniques, while case management and service strategies were recorded from case records. Data were obtained between 1990 and 1992 from a randomly drawn subset of cases employed in the Workload Standards Project of the DYFS. Information was collected by trained coding staff on 102 intake cases, 86 in-home cases, and 51 out-of-home cases. Principal results indicated several distinctive risk profiles. Older children with behavior problems tended to generate a service pattern characterized by a significant number of unsuccessful telephone calls but relatively few unsuccessful field visits by DYFS workers. Economically disadvantaged households with a prior history of abuse and difficulty in recognizing a problem existed were characterized by a high number of unsuccessful field contacts and less supervisory involvement. Households where the parent or caretaker was not employed appeared to trigger a service pattern indicative of higher service intensity and duration, higher supervisory involvement, and more telephone calls. Implications of the results for workflow management and workload and for future research in the areas of risk assessment and case management are considered. 34 references and 5 figures