Get Full Government Meeting Transcripts, Videos, & Alerts Forever!
Get email alerts on the Research Methods topic
No spam. Unsubscribe anytime.
Researchers outline methods, limits and COVID effects on FRC ROI findings
Summary
Wilder Research described a layered methodology—pre/post, propensity scoring, cohort replication and difference-in-differences—and noted data limits, COVID-era effects on reporting, and conservative attribution choices that shaped the reported ROI.
Wilder Research walked the board through multiple analytical layers they used to test whether Scott County Family Resource Centers reduced child-protection involvement and to estimate avoided system costs.
Justin Hollis described the approach: "So our pre post analysis was a very simple before and after question." He said Wilder complemented that with propensity scoring to adjust for differences in families who used services more intensively, cohort replication to test year-to-year consistency, and a difference-in-differences comparison with Carver County to test whether county trends were unique to Scott County.
Presenters noted COVID-era reporting changes — including reduced school-based reports — can affect child-protection assessment rates and must be considered when interpreting trends. They emphasized conservative attribution of avoided events to the FRCs and acknowledged gaps: income and some socioeconomic measures were not tracked in the current file, and the team relied on manual linkage work to match county Salesforce records to SSIS data.
Commissioners requested future analyses that expand tracked outcomes (housing, employment, school readiness) and collect additional participant-level socioeconomic data to better understand differential impacts and improve attribution.
