A Kids at Risk Action project

KARA SIGNAL

The systems are measured. The children are seen.

ContactQuick exit

Methodology

Every number on this site traces to a federal table you can open yourself. This page says exactly which tables, what we did to them, and what we will not do with them.

Sources

AFCARS counts are estimates; states may resubmit and figures can change across vintages. We publish one verified vintage at a time and log vintage changes here. National AFCARS figures are rounded estimates published by the Children's Bureau and are never compared as equals against state counts on this site.

When a published figure, its fiscal year, or its citation changes, the change is logged with its date and its federal table in the corrections log. A site that publishes other institutions' reporting failures has to publish its own.

The not-reported rule

A state that did not report a figure gets a labeled not-reported state everywhere that figure would appear: not a blank cell, not a zero, not an empty bar. The rule is enforced in the database with a constraint and in the code with a type renderers cannot bypass.

The original production test was Massachusetts, which stopped reporting child fatalities through the NCANDS Child File after FFY2021 — Child Maltreatment 2023 showed blank cells for FFY2022 and FFY2023. Child Maltreatment 2024 refills those cells through a separate channel: MA resumed reporting via the Agency File, so the site's Total Fatalities values for MA now read 9 (FFY2022), 11 (FFY2023), 6 (FFY2024). The not-reported rule remains load-bearing; it just no longer has an active state-level example against a currently loaded vintage.

What the federal record does not measure

Every number on this site is a count of children whose maltreatment was reported to a state child protection agency and either investigated or handled through an alternative-response track. That is the entire universe AFCARS and NCANDS see. It is not the same universe as the children being maltreated.

The federal government has actually measured the gap: the National Incidence Study of Child Abuse and Neglect (NIS-4), the fourth in a series conducted at roughly ten-year intervals, sampled maltreatment identified by "sentinels" in schools, hospitals, day-cares, police departments, and public agencies — including cases those sentinels never referred to CPS.

NIS-4 found that CPS investigated 32 percent of the children who experienced Harm Standard maltreatment and 43 percent of children meeting the broader Endangerment Standard. Applying the agencies' own screening criteria to the uninvestigated cases, the study concluded CPS would have investigated 72 percent of the missed Harm Standard children — meaning most of the gap is not caseworker judgment. It is the case never reaching a caseworker.

The gap, in numbers

  • 1,256,000 children under the Harm Standard — 17.1 per 1,000 U.S. children — in study year 2005–06.
  • ~3,000,000 children under the broader Endangerment Standard — roughly 40 per 1,000.
  • Where school staff recognized 52 percent of Harm Standard maltreated children — more than hospitals (11%), police (12%), and the general public (6%) combined.

NIS-4 is the last measurement. The series ran 1979–80 (NIS-1), 1986 (NIS-2), 1993 (NIS-3), and 2005–06 (NIS-4, report issued January 2010). No NIS-5 has been funded. The United States has not measured maltreatment that never reaches CPS in the twenty years since. Every "1 in X children" figure in modern child-welfare coverage traces back to a survey that predates the smartphone.

Source: Fourth National Incidence Study of Child Abuse and Neglect (NIS-4): Report to Congress, Sedlak, A.J., et al., U.S. Department of Health and Human Services, Administration for Children and Families, 2010.

Watchers whose work informs this site

The most useful outside analysis of the federal child welfare data comes from a small group of researchers who publish continuously and cite their tables. Their work does not sit on this site; the site cites it.

  • Marie Cohen — Child Welfare Monitor (analysis) and Lives Cut Short (a project documenting U.S. child maltreatment deaths and the policy choices behind them, sponsored by the American Enterprise Institute). Her CM 2024 analysis is the most rigorous public critique of the federal fatality series and is the source of the counting-changes caveat we carry on /data.
  • NCCPR — the National Coalition for Child Protection Reform's weekly round-up is the source we lean on hardest for the Daily Signal state-level items. When a brief item is labelled "via NCCPR round-up" it is a second-hand source; the round-up carries the underlying links.
  • Safe Passage for Children of Minnesota — their 88-fatality study (funded by former MN Supreme Court Chief Justice Kathleen Blatz, produced with researchers at the U of MN Institute of Child Development and U of St. Thomas law fellows) covers 88 Minnesota maltreatment deaths October 2014 to May 2022 — the model this site treats as the closest state-level equivalent of what CM 2024 does nationally.
  • The Imprint — imprintnews.org is the primary trade publication covering the federal and state child welfare beat; most Daily Signal items link directly to their reporting.