In collaboration with Allegheny County Department of Human Services, this study used a mixed-methods design to document the early service paths of African-American and Caucasian children following child welfare referral, and to identify local characteristics that may contribute to service disparities. Key informant interviews were conducted with caseworkers and supervisors to learn about local characteristics and individual styles of decision-making that may contribute to service disparity. Administrative data from a representative sample of 544 children were also used to examine the contribution of children’s race to the likelihood of case investigation, case opening, and out-of-home placement, controlling for other factors. Stratified analyses by child age were conducted and type of report (Childline or general protective services). African-Americans comprised 13% of the population of Allegheny County in 2006, but nearly 40% of all children who were subjects of a maltreatment report. Preliminary bivariate and multivariate analyses show that child race may be a significant factor in the likelihood of case investigation and case opening, particularly among the very youngest and oldest children referred to child welfare. Among children ages 0 to 2, for example, 77% of African-American children referred were investigated, compared with 69% of White children (p<.01). Among 0 to 2-year-olds investigated, 73% of African-American children received on-going child welfare services, compared with 66% of White children (p<.05). Analysis of the qualitative interviews suggest that when information about risks of child maltreatment is ambiguous or contradictory, caseworkers use several internal processes to arrive at decisions, which may introduce bias. The extent to which these internal processes can become "transparent" may help to reduce decisions that contribute to disproportionality early in child welfare case decision-making. The final results of this study will be available in Spring 2009. Potentially,findings from this study will be used to create strategies at the state- and county-level to address racial disproportionality in child welfare.