How Big Data can help save endangered kids
I do not know who to blame, but I do want people to be held accountable . . . Talk is cheap. We need some actual action.
That was the Rev. Mark V.C. Taylor of the Church of the Open Door in Brooklyn delivering the eulogy last week at the funeral of 6-year-old Zymere Perkins.
Zymere, who allegedly was beaten to death by his mother and her boyfriend in September, is only the latest in a long string of children who have been failed by the adults around them and then by the Administration for Children’s Services.
No one has publicly revealed the extent of the involvement of ACS with the boy, but as far back as April, he was interviewed about various injuries by the police, the Manhattan District Attorney’s Office and child services. As Mayor de Blasio said, “I think what I can safely say is that there were warning signs.” No kidding.
But there might be a solution: Big Data.
According to Richard Gelles, dean of the School of Social Policy and Practice at the University of Pennsylvania, “Even the state-of-the-art assessment tools being used in New York are no better at predicting risk for a child than if you flipped a coin.” Gelles says social workers using “clinical judgment” and their own “expertise” to determine which children should be removed from their homes is “simply inadequate.”
His book “Out of Harm’s Way: Creating an Effective Child Welfare System,” will be released in the spring, and it’s the result of decades studying the problems with these bureaucracies. He has observations the mayor and his advisers should heed.
First, child-welfare agencies are confused about who their “clients” are. They think they’re supposed to be serving parents, not children. So when parents have drug problems, for instance, they try to get the parents into drug-treatment programs.
There’s nothing wrong with that, of course, but it’s not their job. Their job is keeping kids away from parents whose judgment is impaired by drugs.
Second, Gelles argues the agencies charged with maintaining children’s safety are often “siloed” — child services, the police, the District Attorney’s Office and the schools often have very little sense of how much contact the other agencies have had with a child. So it’s hard to determine even the number and severity of the incidents a particular child experiences.
Finally, Gelles notes the failure of existing policy. In the aftermath of high-profile failures, the state has instituted harsher punishment for abusers and established an abuse-reporting hotline. There’s also often a cry for more money and staffers and smaller caseloads for workers. But Gelles has seen no evidence these work.
The answer, says Gelles, is algorithms. Just as insurance companies use them to predict risk, there are ways to measure the risk for a child remaining in a home where there’s evidence of abuse. For instance, he says, social workers tend to underestimate the added risk of the presence of a mother’s boyfriend in the home.
In the past six months, Los Angeles County (the largest child-welfare agency in the world) and Allegheny County in Pennsylvania have both started using this approach. Emily Putnam-Hornstein, a director for the Children’s Data Network at the University of Southern California, helped to develop Allegheny County’s model.
Speaking to PBS Newshour in the spring, she explained, “We have 6 million children [nationwide] reported for abuse or neglect, and how you make triaging decisions early on absolutely impacts outcomes for that child and family.”
Even before these districts tried it out, though, IBM ran a test on data they already had. The algorithm predicted which children were going to be abused again with 90 percent accuracy.
Many will be skeptical at the idea of turning over life-altering decisions to a computer. Aren’t unhappy families, to borrow a phrase, each unhappy in their own way? How will a computer know the difference? We can’t know for sure. But the process is already an informed guessing game; data will just make officials more informed.
One lesson many have taken from these tragic cases is that the state isn’t very good at caring for children. Indeed, there’s no substitute for a stable, married, two-parent family. Sadly, there are communities where such households are a rarity today.
But we can’t simply throw our hands up. There has to be a better way, and now, it looks like there is.
Naomi Schaefer Riley is a senior fellow at the Independent Women’s Forum.Read the full article here