I have enough verified material. Writing the report.

---

The 30-Hour Fact-Finder: How a $477 Million Volunteer Network Became Family Court's Star Witness — With No Outcome Data, No Public Report Card, and Absolute Immunity From Suit

In roughly 900 courthouses across 48 states, the person who tells a judge whether a child should go home is often not a lawyer, not a social worker, and not a state employee. They are a volunteer with 30 hours of training, no license, and — under a doctrine adopted by federal and state appellate courts nationwide — absolute immunity from being sued for what they recommend. The Court Appointed Special Advocate model draws federal Victims of Child Abuse Act money through the Justice Department, state general funds, federal victim-assistance dollars, and court set-asides, for a network whose affiliates reported more than $477 million in combined revenue in 2021, a 57 percent jump in five years. Judges accept these volunteers' recommendations in full in 61 percent of cases and almost in full in another 23 percent. And the single largest rigorous study ever conducted — 31,754 Texas foster children — found that children assigned a CASA had significantly lower odds of achieving permanency and lower odds of being reunified with their families than children who got no advocate at all.

That last sentence is the whole problem. Nearly fifty years and several billion public dollars into the CASA experiment, the federal government does not collect the data that would settle whether it works, the national association's quality reviews are self-administered and unpublished, and the only enforcement action the Justice Department has ever taken against the national body was over a sofa.

How the Model Actually Works

The CASA model was born in 1977 in Seattle, when King County Superior Court Judge David Soukup, uneasy about making removal decisions on thin records, began recruiting citizen volunteers to independently investigate children's cases. Congress blessed it in the Victims of Child Abuse Act of 1990, codified at 34 U.S.C. ch. 203, which authorized a national CASA program administered by the Justice Department's Office of Juvenile Justice and Delinquency Prevention. The Victims of Child Abuse Act Reauthorization Act of 2018 (Pub. L. 115-424) extended it.

The National CASA/GAL Association for Children (EIN 91-1255818, Seattle) sets the standards. Its central requirement is a minimum of 30 hours of pre-service training covering courtroom role, child development, trauma-informed care, and bias awareness. There is no degree requirement, no licensure, no continuing-education board, and no professional body that can strip a credential.

What those 30 hours buy is extraordinary authority. In many states the CASA volunteer is the guardian ad litem — a party or party-equivalent with a statutory right to review confidential child-welfare records, medical files, and school records; to interview the child, the parents, the foster parents, and the caseworker; to file a written report with the court; and to recommend on removal, placement, services, visitation, and termination of parental rights. Texas Family Code Chapter 107 expressly authorizes courts to appoint "a charitable organization composed of volunteer advocates" as the child's guardian ad litem.

Judges listen. In the national evaluation that remains the field's benchmark — 1,994 cases across 25 CASA programs — 61.0 percent of cases saw judges accept 100 percent of the volunteer's recommendations, and in another 22.6 percent judges accepted between 76 and 99 percent. In more than four of five cases, all or nearly all of what the volunteer recommended became the court's order. A National CASA judicial survey found 97 percent of judges agreeing that children and families are better served because of volunteer involvement.

There is a defensible reason judges rely so heavily on CASAs: the same evaluation found judges assign volunteers their hardest cases — unstable placements, conflicting information, extreme neglect, physical and sexual abuse. That selection effect is real, and it complicates every outcome comparison in the literature. It is also precisely why the absence of a randomized trial in dependency court, after five decades, is not an academic footnote. It is the scandal.

The Money: $477 Million, and Nobody Is Adding It Up

The funding runs through at least five channels, and no single agency reconciles them.

Federal (DOJ/OJJDP). Congress has funded the CASA program at roughly $12 million a year — $12 million in FY2018, FY2019 and FY2020, $12.5 million in FY2021 — despite twice appropriating $15 million historically. USASpending records show the Justice Department awarding National CASA $11,000,000 under CFDA 16.547, plus $3,880,608 under 11.555 and $500,000 under 16.582 in a single earlier award cycle. National CASA takes the federal award and re-grants it in subawards.

The national association's own books. Form 990 filings show revenue of $13.5M (2019), $13.07M (2020), $17.01M (2021), $16.99M (2022) — and then a collapse to $8.96M in 2023, against $9.76M in expenses. Headcount went from 53 employees to 66 across the same window. Aggregate officer, director and trustee compensation ran from $526,323 in the earliest reported year to $637,703 most recently.

Subawards to the field. National CASA's Schedule I shows the pass-through: Connecticut CASA $149,099; CASA Mississippi $103,600; Richland County CASA (SC) $85,973; CASA of Philadelphia County $85,141; CASA of New Hampshire $80,000; West Virginia CASA $75,900; Seventh Circuit CASA (SD) $75,490; CASA of Missoula $74,600; Snohomish County CASA $74,120; Maryland CASA $68,000 — roughly $2.82 million in traceable grants made.

States. Texas commits about $15,964,000 per year through the Health and Human Services Commission — $31.9 million for the 2024–25 biennium — for a program that served 17,864 children in FY2024, an 11.4 percent decline from the prior year. Virginia budgets roughly $1.6 million annually in general funds through the Department of Criminal Justice Services, and for FY2026 has $1,533,840 in state general funds plus $1,499,490 in federal Victims of Crime Act money flowing to local CASA programs.

Court set-asides. Section 438 of Title IV-B, subpart 2 of the Social Security Act sets aside $40 million annually for the State Court Improvement Program — formula grants to the highest court in every state plus $1 million a year for tribal courts. CIP money has been used for CASA volunteer training, among other purposes, which means Title IV-E/IV-B child welfare dollars underwrite the advocate whose recommendations then drive IV-E placement decisions.

Add it up at the affiliate level and the number is genuinely large: more than $477 million in combined 2021 revenue across roughly 900 independent CASA affiliates, up 57 percent since 2016 — a figure that rivals what states like Colorado, Wisconsin, Kansas, Alabama, Oklahoma and Nebraska spend on their entire child welfare systems. Private philanthropy tops it off: Jewelers for Children gave National CASA $225,000 in 2022 and $200,000 in 2023; Fidelity Charitable, Schwab Charitable, the American Endowment Foundation, the American Online Giving Foundation and the Tides Foundation all appear in the Schedule I record.

"Back-Order Sofa": What Federal Oversight Actually Caught

In 2023 the Justice Department designated National CASA a high-risk grantee — a status reserved for documented unsatisfactory performance, financial instability, or internal control deficiencies — and froze its largest funding source. The underlying federal financial review questioned more than $2.7 million in expenditures and found "inadequate" accounting practices and high turnover in the chief financial officer role.

The itemization is worth reading closely, because it defines what the federal government was actually looking at: $1.14 million in office lease and storage payments across Seattle, Atlanta and Washington, D.C.; roughly $765,000 to a Las Vegas public relations firm for a "2021 Media Campaign"; more than $298,000 for "recruitment fees and legal services for trademarks"; and $4,099.81 for a "Back-order Sofa." After negotiation, $273,132.68 was ruled unallowable and repaid. By October 2024 the organization was in its nineteenth month as a high-risk grantee.

Then, on April 22, 2025, DOJ issued more than 360 grant termination notices — roughly 365 awards valued at about $811 million at the time of award — and National CASA's three grants were among them, terminated on the ground that they "no longer effectuate the program goals or agency priorities." Effective April 24, National CASA suspended all OJJDP-funded services and subawards, warning that more than 200,000 children in abuse and neglect proceedings would be affected. The association appealed and the grants were reinstated after DEI-related terms were removed from the awards.

Here is the point that matters. Every one of those federal actions — the freeze, the clawback, the termination, the reinstatement — turned on accounting, terminology, and administrative priorities. Not one turned on whether a single child was better off. The federal machinery has a trigger for a mis-booked sofa. It has no trigger for a volunteer who never met the child whose fate they recommended.

The Evidence: What Four Decades of Research Found — and Didn't

The 2004 Caliber Associates national evaluation, prepared for OJJDP and commissioned by National CASA itself, remains the most-cited study. Its findings are not the ones the brochures quote. Children assigned a CASA were more likely to be placed out of home and less likely to be reunified with family or placed with kin. In the closed-case subsample, 89 percent of children with a CASA volunteer were placed in out-of-home care versus 18 percent of children without one — a covariate-adjusted comparison significant at p<.001. Children without CASAs were no more likely to be re-maltreated, and stayed in the system about the same amount of time.

One correction to the common critique is owed here: Caliber found that CASA-assigned children and their parents received more services, not fewer. That cuts in the model's favor and should be stated plainly. The problem is that more services did not translate into more permanency or less re-maltreatment.

The most rigorous modern study is Osborne, Warner-Doe, LeClear & Sexton, "The Effect of CASA on Child Welfare Permanency Outcomes," Child Maltreatment (2020) — an intent-to-treat design using inverse probability weighting on 31,754 Texas foster children to address selection bias. It found children appointed a CASA had significantly lower odds of achieving permanency, lower odds of reunification, lower odds of permanent kin guardianship, and greater odds of adoption if not reunified. CASA appointment was associated with delays in family reunification. Dallas CASA's public response was to publish a rebuttal titled "Flawed Study Fails to Measure CASA's Impact."

Ohio's own commissioned evaluation — conducted by the Ohio Colleges of Medicine Government Resource Center at Ohio State, released August 30, 2023, covering roughly 8,000 children — produced the most uncomfortable finding in the literature. CASA appointment correlated with more regular medical appointments. But in rural counties, appointment of a volunteer reduced time in out-of-home placement for White children and not for children of color, and children of color with volunteers moved between foster homes more frequently than their White peers. A program designed to protect children delivered its benefit along a color line, in a study its own state association paid for.

Lawson & Berrick, "Establishing CASA as an Evidence-Based Practice," Journal of Evidence-Based Social Work 10:321–337 (2013), reviewed the effectiveness claims and concluded they are "not fully supported by the literature and should be viewed with caution" — noting, for instance, that on foster care reentry, two of the three studies cited as supporting evidence had no statistically significant results. A Campbell Collaboration systematic review protocol was registered in 2015. More than a decade later the field still has no completed authoritative synthesis.

The first true randomized controlled trial — Ryan, "From Advocacy to Outcomes: A Randomized Controlled Trial of CASA in Juvenile Justice," Juvenile and Family Court Journal (2025) — found real benefits: reduced recidivism and improved academic self-efficacy, with no effect on emotional or social self-efficacy. It is genuine evidence and it deserves credit. It is also a study of adolescents on probation, not of removal and permanency decisions in dependency court. The question at the heart of the model — does a 30-hour volunteer's recommendation make removal decisions better or worse? — has never been randomized.

And there is a structural reason why. Washington State's own CASA evaluation feasibility study found that "information about time frames and case progression is not routinely collected by CASA/GAL programs, and there is no standard method by which volunteers or programs keep records of case activities and outcomes."

The Race Gap Is in the Hours

The disparity that is best documented is not who gets an advocate — it is what the advocate does once assigned. Caliber found CASA volunteers spent a mean of 2.67 hours per month on cases involving African American children versus 4.30 hours on cases involving children of other races. That is a 38 percent gap in attention, on cases where the volunteer's report substantially determines the outcome.

The demographic mismatch is stark and stable. Caliber-era volunteers were 90 percent white and 8 percent African American, serving a child population that was 48 percent white and 36 percent Black or biracial. A 2014 national survey put volunteers at 81 percent white, 82 percent female, 69 percent college-educated, deployed among families overwhelmingly poor and disproportionately non-white — the core empirical claim of Amy Mulzer and Tara Urs's law review article However Kindly Intentioned: Structural Racism and Volunteer CASA Programs, 20 CUNY L. Rev. 23 (2016). Local numbers echo it: CASA for Douglas County, Nebraska, served 485 children in 2022–23, 40 percent of them Black, with 5 percent of its volunteer advocates identifying as Black.

On the narrower claim that Black children are less likely to be appointed an advocate at all, the public record is suggestive but thinner than the hours data — county-level demographic comparisons like Fairfax's are confounded by underlying disproportionality in who enters care. The honest statement is that the disparity in advocate effort and advocate benefit is documented in two independent evaluations a generation apart, and that no federal agency requires any program to report appointment rates by race.

Immunity: The Door That Closes Last

If a CASA volunteer's investigation is negligent, incomplete, or simply wrong, the tort system is generally unavailable. Courts have extended absolute quasi-judicial immunity to guardians ad litem on the theory that they perform tasks — investigating, reporting, testifying, recommending — so integral to the judicial function that they are treated as an arm of the judge.

Fleming v. Asbill, 326 S.C. 49, 483 S.E.2d 751 (1997), held that guardians ad litem are entitled to absolute quasi-judicial immunity for acts within the scope of appointment (its federal companion is 42 F.3d 886 (4th Cir. 1994)). The Eighth Circuit has affirmed dismissal on the same ground, holding guardians, therapists and court-appointed attorneys absolutely immune from damage claims tied to their court testimony. Kentucky's Court of Appeals published a decision reaching the same result in 2021; the Illinois Supreme Court has granted GALs quasi-judicial immunity as well. Legislatures have codified it: Texas Family Code § 107.009 provides that a guardian ad litem, attorney ad litem or amicus attorney "is not liable for civil damages arising from an action taken, a recommendation made, or an opinion given" in that capacity, subject only to an exception for conscious indifference or reckless disregard.

The Fleming court listed the safeguards that supposedly make immunity tolerable: cross-examination, oversight by the appointing court, a party's ability to move for the guardian's termination, the court's power to reject recommendations, and appellate review. Every one of those depends on a parent having a lawyer with time, investigative resources, and an appellate record. In dependency court, the parent is typically indigent and represented by appointed counsel carrying a caseload that makes cross-examining a volunteer's methodology a luxury. The safeguards are real on paper and largely theoretical in a Tuesday-morning docket.

When Programs Fail

Kentucky. An audit of the CASA program serving Carroll, Grant and Owen counties found the program allowed volunteers to work cases without required background checks or adequate training. Reviewers documented "numerous instances of egregious national CASA standards and Kentucky Revised Statute violations" and ordered all volunteer and staff case activity to cease immediately. On January 18, the local board dissolved the program. Note what did not happen: no state lost funding, no federal award was conditioned, no children's cases were systematically reopened. The program closed itself.

Travis County, Texas. CASA volunteer Karen Rose Slapik was charged with endangering a child after admitting to police she gave alcohol to a 12-year-old and a 15-year-old assigned to her and let the older girl drive her car. CASA of Travis County discharged her and notified police.

Santa Cruz County, California. A CASA volunteer active in the program since 2009 was arrested on suspicion of child sexual abuse, annoying or molesting a child under 18, possession of child sexual abuse material, and indecent exposure — conduct connected to his role in the program. CASA requires FBI, sex-offender registry and criminal history screening. The screening is a point-in-time check; it does not surface what a volunteer does over sixteen years of unsupervised access to vulnerable children.

The Accountability Gap

Four bodies could be watching, and each one is watching something else.

OJJDP monitors grant compliance — ledgers, allowable costs, cost allocation. Its most forceful intervention in the program's history produced a $273,132.68 repayment over office leases and a sofa.

National CASA administers quality assurance through a Standards Self-Assessment and Indicators of Compliance completed every four years, scored by an outside reviewer who returns a report to the program in ten to twelve weeks. It is a self-assessment, the reports are not published, and there is no public registry showing which of the roughly 900 programs passed, which failed, or which were remediated.

HHS runs AFCARS, the federally mandated case-level data system covering every child in foster care under Title IV-B/IV-E. It does not publish CASA appointment as a linked outcome variable. There is no federal dataset that lets anyone ask whether children with advocates do better.

The courts are the theoretical backstop, and immunity doctrine has removed damages liability as a corrective.

The result is a system in which a program can be shut down for background-check violations, a national body can be frozen for accounting failures, and a volunteer can be arrested for abusing the child they were appointed to protect — and none of it is connected to any measurement of whether the model helps children get home.

Why It Matters, and What Would Fix It

Removing a child from their parents is the most consequential thing a state does to a family short of imprisonment. If the Osborne finding is right — that CASA appointment is associated with lower odds of reunification and lower odds of kin guardianship in a sample of 31,754 children — then the intervention is, at the margin, keeping children in foster care longer. If the Ohio finding is right, it is doing so unevenly by race. Neither finding is conclusive, because the selection problem is genuine and judges really do assign advocates to the hardest cases. But that is an argument for a trial, not for another four decades of testimonials.

Five things would resolve this, and none of them require abolishing anything:

1. Make CASA appointment a reportable AFCARS element. One data field, linked to existing case-level records, would let researchers answer the permanency question definitively within two reporting cycles.
2. Condition VOCAA funds on publication. Require every program drawing federal money to publish its quality-assurance result, its volunteer-to-child contact compliance rate, and its appointment and advocate-hours data broken out by race.
3. Fund one adequately powered multi-site randomized trial in dependency court. Ryan's 2025 RCT proves it can be done ethically in juvenile justice. The dependency version is the study that has never been run.
4. Buy the alternative that already has evidence. Since 2019 the federal Children's Bureau has permitted states to claim Title IV-E administrative funds for legal representation of parents and children. Peer-reviewed evaluation of interdisciplinary parent representation in New York City found children exited foster care substantially faster with no increase in re-entry or safety incidents. That is a better-evidenced use of the same federal dollar.
5. Narrow immunity to testimony. Absolute immunity for what an advocate says on the stand is defensible. Absolute immunity for failing to conduct the investigation the statute requires is not. Pair it with program-level liability so the organization, not the volunteer, carries the risk.

The CASA network was built on a decent instinct — that no child should stand alone in front of a judge. Tens of thousands of volunteers give real time to real children for nothing. But good intentions are not an evidence base, and a self-graded report card is not oversight. A model that directs half a billion dollars a year and shapes the outcome of four out of five cases it touches should be able to answer one question: are the children better off? Nearly fifty years in, no agency in the United States can.

---

Sources: - Evaluation of CASA Representation, Final Report — Caliber Associates (OJJDP) - The Effect of CASA on Child Welfare Permanency Outcomes — Osborne et al., Child Maltreatment (2020) · PubMed - Ohio CASA Evaluation Final Report, Aug. 30, 2023 — OSU Government Resource Center - From Advocacy to Outcomes: An RCT of CASA in Juvenile Justice — Ryan, Juvenile and Family Court Journal (2025) - Establishing CASA as an Evidence-Based Practice — Lawson & Berrick - However Kindly Intentioned: Structural Racism and Volunteer CASA Programs — Mulzer & Urs, 20 CUNY L. Rev. 23 - Spending on CASA for Foster Kids Surges; Evaluation Lags — The Imprint - After Fed Scrutiny, National CASA Repays Hundreds of Thousands — The Imprint - DOJ Freezes Funding for National Child Advocates Group — The Imprint - Federal Grants Restored to CASA-GAL After Removal of DEI Terms — The Imprint - National CASA/GAL federal funding terminated · CASA Appropriation · Judicial Survey Results - OJJDP — National CASA/GAL Membership, Accreditation, Training and TA Program - Victims of Child Abuse Act Reauthorization Act of 2018, Pub. L. 115-424 · 34 U.S.C. Ch. 203 - State Court Improvement Program — ACF Children's Bureau - National CASA Quality Assurance System FAQ · Standards for Local CASA/GAL Programs - CASA Evaluation Feasibility Study — Washington Courts - Fleming v. Asbill — S.C. Supreme Court · 4th Cir., 42 F.3d 886 · Guardians ad litem have absolute immunity — ABA Journal - Texas Family Code Ch. 107 excerpts — Texas CASA · Texas CASA HHSC funding - FY2026 Court Appointed Special Advocate (CASA) — Virginia DCJS - Audit: Former CASA program for Grant, Owen and Carroll Counties broke state laws — WCPO I-Team - CASA volunteer accused of giving foster children alcohol — KXAN · Santa Cruz youth advocate arrested — CBS News - The case against CASA — NCCPR - IRS Form 990 filings, National Court Appointed Special Advocate Association, EIN 91-1255818 (TY2019–TY2023), and USASpending award records — retrieved via PMC CivicOps database.

Two corrections to the framing in the tasking directive, both of which strengthen rather than weaken the piece: Caliber found CASA-assigned children and parents received more services, not fewer — the failure is that more services did not yield more permanency. And the best-documented racial disparity is in advocate hours and benefit (2.67 vs. 4.30 hours/month; Ohio's White-only placement benefit), not conclusively in who gets appointed an advocate; no agency requires appointment-rate reporting by race, which is itself the finding.