Predictive analytics, mathematical equations that forecast where events will likely occur, and the use of data within governmental agencies such as the Texas Department of Family and Protective Services (DFPS) are seen by many as a way to proactively penalize people before they ever do anything wrong.
This misunderstood view has led many people who are uncomfortable with the perceived possibility of misuse to question the place of such methods within the government. However, as DFPS system failures continue to result in the deaths of Texas children the question arises: are you willing to give predictive analytics the chance to save the life of a child?
As Governor Greg Abbott and the Texas legislature look for ways to improve the child welfare system and reduce the rates of child maltreatment, many people do not understand how data can be used to prevent child abuse and neglect.
Texas had 63,781 sustained reports of child maltreatment in 2015. I believe it is imperative for the state government to search for realistic and effective solutions to drastically reduce the rate of child abuse and neglect. When discussing how big data and predictive analytics can be used, many of my friends, family, and peers are wary of giving the government so much power.
Movies like “Minority Report” have provided a picture of how extreme – if fictional – uses of predictive analytics can be a punitive way to strip people of their rights without justification. People worry that Child Protective Services (CPS) will target families with high-risk scores, and those families will be repeatedly hounded by CPS until children are taken from their parents. While the government punishing people prematurely and without cause is something to be cautious about, the realities of how data can work within the child welfare system are intriguing and creative.
A key aspect to consider when looking at the role of data in child welfare is the difference between CPS’ approaches to intervention and prevention. Predictive analytics can be used as a way to stimulate necessary intervention or they can help states designate prevention services to higher risk areas to stop maltreatment and CPS intervention from occurring.
Emily Putnam-Hornstein, an associate professor at USC’s Suzanne Dworak-Peck School of Social Work and director of the Children’s Data Network, described Allegheny County’s use of predictive data in formulating risk scores for families as a means to improve the screening in or out process for child maltreatment reports.
For example, in Allegheny County a call center employee receives a report on a family and, based on the reported details and the predicted risk score, will triage the report to a specific team. The risk scores help the call center employees and management decide whether to send a child service representative to visit the family even if the details reported may not have solely been enough to trigger a visit. The call center employee and manager are privy to the risk score but the representative investigating the case is not. Keeping the generated risk scores from the representatives who actually interact with the families ensures the representative does not have preconceived assumptions about the family they evaluate.
In addition to the use of data to calculate risk scores and aid CPS in its efforts to intervene and protect children from maltreatment, data can also be used as an effective method of identifying which geographic areas are most in need of preventative services.
In the fall of 2016, The Center for Prevention of Child Maltreatment at Cook Children’s Health Care System and Texas Christian University’s Department of Criminal Justice published an intriguing and innovative study of how Risk Terrain Modeling (RTM) was used as a tool to more accurately predict high-risk areas. While the model was initially tested on previous data from Tarrant County, Breanna Anderson, the Program Coordinator for The Center for Prevention of Child Maltreatment led by Cook Children’s, believes that RTM “could be a useful tool for every organization, state, and nation to use to effectively target services.” These prevention services could provide high-risk families with the skills and support they need to raise their children in a healthy environment.
While the thought of having the government look at data that predicts future behavior is worrying to the public, a more detailed examination of the application will hopefully relieve specific fears. This is not “Minority Report,” we are not Tom Cruise, and the only use for predictive analytics is not preemptively punishing citizens without cause.
Predictive analytics are a way to save children who have few ways to protect themselves. I understand the hesitation, but I encourage everyone to take time to consider the options, speak with professionals about the intricacies of utilizing data, and learn more about what your state is currently doing or proposing.
If you then feel that these methods could help children, contact your representatives about how data is and can be used in the prevention and intervention of child maltreatment. We all have a voice, and it is our responsibility to use our voice to speak for children who too often cannot speak for themselves.
Bethany McKee is a Master of Public Policy student at USC’s Sol Price School of Public Policy with interests including education and social justice. She currently serves on the board of A Spacious Place, Inc., an Austin-area nonprofit that provides creativity classes, clubs and camps to underserved communities. She wrote this story for the Media for Policy Change course at USC.