Project description:

Our investigation went beyond the prevailing narratives about child welfare investigations — that state agencies either do too much or too little to keep children safe. Instead, we revealed a system that deprives people of their rights in ways that neither protect children nor their families. We tracked the sequence of a child welfare case, examining how authorities wielded their power to investigate and separate families — beginning with the initial, usually anonymous, tip that a child might be in danger. Our investigation was fueled by an unprecedented analysis of tens of millions of child welfare records from the past decade.

Impact reached:

In the wake of our story in Arizona, the office of the state’s incoming governor said it had read our work and wanted to take the agency in a new direction. She chose a new director of the state’s CPS department: a Black community advocate whom we had highlighted in our story.

In New York, both the City Council and the state Assembly appear poised to pass bills creating a “Miranda warning” notifying families being investigated by CPS of their rights. This comes after our story about how the city conducts tens of thousands of warrantless home searches every year. Our report had highlighted the Miranda idea as a possible solution.

In response to ProPublica and NBC News’ findings on mandatory reporting, a bipartisan group of state lawmakers in Pennsylvania has called for the legislature to conduct a study of the 2014 child welfare law that we highlighted. In West Virginia, two state legislators said their staffs would examine the state’s practice of terminating parental rights in light of our story there. In response to our story in Illinois, two lawmakers in the state called on their governor to improve access to mental health and substance abuse treatment to ensure that families repeatedly investigated by the state’s child welfare agency can access help.

Our project could also prompt the federal government to pass reforms to the child welfare policies that our series examined: Members of Congress are planning to introduce a bill this year that would end the strict federal timeline for the termination of parental rights. They were shocked by our findings — including a state-by-state analysis of termination timelines that no other institution has conducted — and agreed that such realities could help shift attitudes on the child welfare system and create bipartisan momentum for change.

Techniques/technologies used:

We filed requests for data from the National Data Archive on Child Abuse and Neglect for their National Child Abuse Data System (NCANDS) and Adoption and Foster Care Analysis and Reporting System (AFCARS). The NCANDS data required a research proposal documenting how we would keep the information secure and approval from an institutional review board.

Both databases are incredibly large (millions of entries per year) and messy, so over the course of a year, we developed a careful methodology for analyzing them. A single child can appear many times in the data, so we employed multiple large queries — written and executed in R — to pull relevant information from different parts of each database to create as accurate a list of children and cases as possible. Even still, we made sure to be as conservative as possible in interpreting our analysis results, especially because there were differences in how certain fields were tracked among states. We also shared our methodology and analysis results with multiple experts familiar with both sets of data as well as the data archive itself who confirmed the validity of our methods.

Our reporting in Illinois used data obtained from that state’s child welfare department through a public records request. This data allowed us to track the same family over time, which isn’t possible with NDACAN. The resulting story drew attention to thousands of Illinois families who cycle through multiple cases, centering on one rural family whom authorities investigated 10 times without successfully stabilizing them.

Context about the project:

To access the NCANDS data used in this project, we had to seek IRB approval, which required a written project proposal and $1,125 fee.

Politically, child welfare is a fraught topic. Stories typically focus on outlier events, such as cases of extreme child abuse or neglect or stories of innocent parents who were wronged by the system. But our data analysis and reporting told a different story – one in which millions of families became ensnared by a system that has much the same power and disproportionality of the criminal justice system without the same constitutional protections. This required uncovering the trends through data analysis and featuring parents who may not be perfect, but for whom the punishment didn’t appear proportional to what they were accused of. Reporting this story was particularly challenging because access to documents is tightly guarded in nearly every state and child protective services workers and judges don’t typically comment on them. This required meticulous work on the part of the reporters to comb through what was publicly available and fact check what our sources were telling us.

What can other journalists learn from this project?

The greatest strength of this project is that it documented systemic issues in a topic in which most investigations typically focus on outlier events. For example, most assume child protective services only deal with cases of extreme abuse, but our analysis found that the majority of cases do not involve those allegations, and very few are substantiated. By zooming out and cataloging the outcomes and disparities of the system, we were able to show the very real harm it has on the majority of people it comes in contact with, and the ineffectiveness of supposedly “common sense” policies, such as mandatory reporting.

From a data perspective, the project’s success in uncovering meaningful findings on child welfare, a traditionally difficult and opaque subject, demonstrated that there are many opportunities to do original and impactful reporting on traditionally under-covered or poorly documented subject areas.

This project also showed that there is a lot to be gained in telling the stories of people who may not be traditional “perfect” victims. In each story, we pushed ourselves to feature parents who were representative of the trends we found in our reporting, regardless of how spotless their records were.

Project links:

https://www.propublica.org/article/for-black-families-in-phoenix-child-welfare-investigations-are-constant-threat

https://www.propublica.org/article/six-months-or-less-parents-lose-kids-forever

https://www.propublica.org/article/how-we-analyzed-child-welfare-investigation-data

https://www.propublica.org/article/dcfs-illinois-investigations-child-welfare

https://www.propublica.org/article/child-welfare-search-seizure-without-warrants

https://www.propublica.org/article/mandatory-reporting-strains-systems-punishes-poor-families

https://www.propublica.org/article/arizona-matthew-stewart-child-safety-hobbs