Battle Over Research Findings Amid Rocky Transition in Kansas
The transition of leadership in child welfare agencies is rarely a newsmaking event. Not so this year in Kansas, where the shift from Gov. Jeff Colyer (R) to Laura Kelly (D) has been marked by a pointed research dispute about welfare benefits and foster care, cancelled grants and accusations of obstruction.
Kelly has named Laura Howard, director of the Public Management Center in the School of Public Affairs and Administration at the University of Kansas, to serve as interim director of the Department of Children and Families (DCF). She will replace Colyer’s DCF boss, Gina Meier-Hummel, on January 14.
At the same time she announced the hire, Kelly also said she’s putting a hold on a slate of grants made by DCF this year and returning DCF procurement to a contract-based system. Meier-Hummel oversaw the awarding of the grant in 2018.
“Because the grants weren’t considered by a separate entity, the evaluations weren’t neutral,” said Lori Ross, CEO of FosterAdopt Connect, in an interview with the Wichita Eagle. “It was all done inside DCF. … When an agency awards its own grants, it may select who they want to do business with and not who is best for Kansas’ kids and families.”
In discussing the return to contract, Kelly made a point of telling reporters that she felt DCF had not been helpful in the transition.
“Despite our best efforts during the transition, accurate and forthright information from current DCF leadership was hard to come by,” Kelly told the Eagle.
DCF pushed back against researchers’ claim that the state’s recent limitations on cash assistance to poor families is an underlying cause of the state’s increasing foster care totals.
About a year ago, University of Kansas researchers disseminated preliminary findings of a study, funded by the Centers on Disease Control and Prevention, asserting a causal connection between cash assistance to families and negative child welfare outcomes.
The report, a first installment in a broader project about the social safety net and child neglect, asserted that limitations on payments from the Temporary Assistance for Needy Families (TANF) program led to increased child victimization and placements into foster care.
In April of 2018, at a presentation about the Kansas child welfare system, the authors prepared a PowerPoint presentation singling out the state on this issue.
Earlier this week DCF leadership fired back with an independent evaluation that refutes the study and questions the integrity of its methodology.
“Following the robust review, DCF found no correlation existed between TANF time limits and foster care,” DCF noted in a statement issued this week.
The evaluation, led by University of Maryland professor and renowned child welfare researcher Douglas Besharov, questions the research and evaluation methods used by University of Kansas researchers Donna Ginther and Michelle Johnson-Motoyama to assess the child welfare impact of restrictions to Temporary Assistance for Needy Families (TANF).
“You want the best possible evidence to shape the policy changes you might make,” said Douglas Besharov, in an interview with The Imprint. “The publication we were asked to review doesn’t meet that standard.”
In the evaluation, Besharov and research partners criticize the use of incomplete data and analysis, incomplete and shifting variables concerning TANF and other program changes, other changes in child welfare policy and practices that are not considered, increases in opioid and other substance use during the time frame, and even the interpretation of the results by the authors.
“Professor Besharov’s study plainly illustrates we cannot simply rely on one study to change public policy or law, but instead we must be diligent about reviewing public policies, data and outcomes to ensure we are doing the right things for Kansas children and families,” said Meier-Hummel, in the release. “My team continues to look at this matter internally.”
There are many variables when considering child welfare, welfare reform and other public services and policies, but can be difficult for researchers to discern, Besharov said. Sometimes simple questions, like how teen pregnancy rates have been impacted by welfare reform or how does TANF benefit kinship families, can be much more complex, Besharov said, and the research must be robust and consider all the factors, before crafting policies to address the problems.
“Those are all giant questions and they have to be addressed rigorously and carefully,” Besharov said. “Let’s make sure the study or data is correct before we create policy around it.”
The evaluation also notes that the University of Kansas researchers did not respond to requests for additional information.
Johnson-Motoyama did respond to The Imprint, and called Besharov’s evaluation “incredibly offensive” and partisan.
“We welcome a series of rigorous studies on the subject and hope under new leadership that DCF will embark on an examination of their own data to inform evidence-based policy making,” Johnson-Motoyama said. “The dismissive and authoritative arguments by Drs. Besharov and [Neil] Gilbert have offered nothing constructive in this process; rather, they accepted $24,000 in Kansas taxpayer dollars to advance an ideology.”
Johnson-Motoyama conceded that “our causal analysis is not perfect,” but strongly disagreed “that our interpretation of results is somehow problematic.” She also dismissed the evaluation’s criticism of their work, including additional data as the research evolved.
“With Dr. Besharov and Dr. Gilbert’s logic, no scholar can change or improve on their analysis, even when better or more illustrative data becomes available,” she said. “Even with new data, our results are remarkably consistent.”
The University of Kansas study is part of a three-year, $1.05 million grant from the Centers for Disease Control to study how changes to economic and social safety net policies affect child neglect. The full study will be published in the National Bureau of Economic Research later this year.
John Kelly contributed to this article.