Predictive analytics uses data to discover patterns and make predictions about future outcomes. When used responsibly, predictive analytics can enhance decision-making about child welfare services and child protection interventions. The strategy can also identify systemic racial biases by identifying variations in responses to families based on race. Predictive analytics must be used with caution, however, since without ethical oversight and the careful application of data, child welfare agencies risk interpreting results inaccurately or developing a misguided tool. This could lead to increased attention from the child protection system for certain families, which could result in oversurveillance and unnecessary child welfare intervention.

There are steps agencies should take to ensure they are creating and using predictive analytics tools in a way that will produce accurate, ethical results. Agencies should use high-quality data with sufficient depth and breadth to cover an array of potential risks and protective factors and provide proper training and guidelines for appropriate use of the data. Through careful and intentional application of predictive analytics as one component of a larger strategy, child welfare leaders can improve child and family well-being. The resources below provide more information about predictive analytics in child welfare and about challenges and opportunities the technology presents for the field. 

Predictive Analytics in Child Welfare
U.S. Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation
Offers information on how predictive analytics is being used in child welfare and presents a series of resources that discuss challenges and opportunities for this technology. 

Predictive Analytics in Child Welfare: An Overview of Current Initiatives and Ethical Frameworks to Inform Equitable Policy and Practice [Video]
Lanier & Gibbs (2021) 
University of North Carolina at Chapel Hill
Presents an overview of the use of predictive analytics in child welfare, including algorithm-based decision-making, and how to apply an ethical framework while using data to inform child welfare policy. 

Predictive Analytics in Child Welfare — Benefits and Challenges 
Jackson (2018) 
Social Work Today, 18(10)
Analyzes the benefits and challenges of predictive risk-modeling tools and their use in child welfare. The article discusses their effectiveness, concerns about bias, obstacles to widespread use, and more. 

Predictive Risk Modeling – Analytics and Child Welfare: An Examination [Webinar]  
New York University, McSilver Institute for Poverty Policy and Research & Fordham University, Graduate School of Social Service (2021)
Explores the implementation of predictive analytics in child welfare. The webinar shares viewpoints from proponents, who say these tools improve child protection and reveal racial disparities, and critics, who say these tools perpetuate those same disparities.

(Back to Top)

Considerations for child welfare

Big Problems, Big Solutions, Big Data: A Defense of the Use of Predictive Analytics in Child Welfare
Katz 
The New Social Worker
Provides examples of jurisdictions using predictive analytics in child welfare and argues that predictive analytics is a tool that can be used to solve systemic problems within the field. 

Considerations for Implementing Predictive Analytics in Child Welfare (PDF - 956 KB) 
Casey Family Programs (2018) 
Presents information for jurisdictions interested in using predictive analytics in child welfare and reviews issues associated with implementation as well as issues to consider when using this tool. 

Making the Most of Predictive Analytics: Responsive and Innovative Uses in Child Welfare Policy and Practice (PDF - 692 KB)
Chadwick Center for Children and Families & Chapin Hall at the University of Chicago (2018)
Presents the benefits and pitfalls of predictive analytics and examines principles for responsible use.

Predictive Risk Modeling for Child Protection
Centre for Social Data Analytics, Auckland University of Technology, Children's Data Network, & Mathematica (2019)
Discusses how predictive risk modeling could be used to help identify and protect children who are referred to the child welfare system.  

Principles for Predictive Analytics in Child Welfare
Scharenbroch, Park, & Johnson (2017) 
NCCD Children's Research Center
Analyzes how using predictive analytics in child welfare can drive better outcomes for children and families when used within the context of explicitly defined values and principles.

(Back to Top)
 

Challenges and potential risks of predictive analytics

Coding Over the Cracks: Predictive Analytics and Child Protection (PDF - 638 KB) 
Glaberson (2019)
Fordham Urban Law Journal, 46(2)
Explores the risks of using predictive analytics in child welfare decision-making and reviews the potentially biased human processes that go into the creation of these tools. 

Foretelling the Future: A Critical Perspective on the Use of Predictive Analytics in Child Welfare (PDF - 209 KB) 
Capatosto (2017) 
Kirwan Institute for the Study of Race and Ethnicity 
Discusses the use of predictive analytics as a decision-making tool in child welfare and explains concerns about the widespread utilization of this technology. 

Race, Equity, and Ethics: Questions on Child Welfare and Predictive Analytics [Webinar]
Center for the Study of Social Policy (2017) 
Discusses the structural factors that bring families to the attention of child welfare and how agencies should keep these factors in mind when implementing predictive analytics. 

Using Algorithms and Artificial Intelligence in Child Welfare 
Corrigan (2019) 
Reviews the role of algorithms and artificial intelligence in child welfare and presents risks, safety concerns, and challenges about their use. 

Using Artificial Intelligence, Machine Learning, and Predictive Analytics in Decision-Making
Pryce, Yelick, Zhang, & Fields (2018)
Florida Institute for Child Welfare
Examines the use of predictive analytics in child welfare and discusses the impact this technology may have on bias in decision-making.

(Back to Top)