全球智库信息集成服务系统

Predictive modeling and rapid-cycle evaluation— both individually and together—hold significant promise to improve programs in an increasingly fast-paced policy and political environment. 

The authors propose that social service agencies take two actions. First, agency departments with planning and oversight responsibilities should encourage the staff of individual programs to conduct a thorough needs assessment. This assessment should identify where predictive analytics and rapid-cycle evaluation can be used to improve service delivery and program management. The assessment should also evaluate whether the benefits of adopting these tools outweigh the costs, resulting in a recommendation of whether and how these tools should be deployed. Second, federal agencies should take broad steps to promote the use of predictive analytics and rapid-cycle evaluation across multiple programs. These steps include investments in data quality and data linkage, as well as measures to support and promote innovation among agency staff.