Child Protective Services Mistakes Demand Due Process Solutions

Prof. Dwyer’s second essay, “Child Welfare and the State’s Fiduciary Responsibility,” posted November 20, 2018, challenges me, again, to “prove” the problem of wrongful state intervention against families who should neither be investigated nor separated. Dwyer’s second essay takes some new turns, however, and calls out for further clarification, lest our areas of agreement and disagreement become even cloudier. The point of a debate is to illuminate a topic, not interject diversionary disputes. It also is worth pausing to note when we reach the same conclusion from dramatically different premises and priorities.

The focus of my lead essay was on mistakes child protective services (CPS) make in responding to hotline calls too precipitously and too intrusively. Specifically, I focused on the problem of informal family separations under “safety plan agreements” during CPS investigations. Walter Olson helpfully described these separations of families as occurring within a larger legal context—in the “shadow of the law.”

Prof. Dwyer’s second response, like his first, denies that I have identified a real problem. He starts with a discussion of how hotline call investigations operate, introducing what seems, at first blush, to be a useful distinction between “valid” and “founded” allegations. The distinction he offers, fairly enough, is that a “valid” allegation is one that essentially “states a claim” for abuse or neglect, while a “founded” allegation is one with a sufficient evidentiary basis.

Prof. Dwyer’s claim is that I haven’t shown a rampant problem of allegations being “deemed valid when they are not.” But valid allegations are not necessarily allegations that warrant investigation.

Some definite and reliable evidence of maltreatment of a child should be required before CPS embarks on an investigation, given that investigations almost inevitably carry with them seeds of distress and disruption for children and families alike. Creating such a threshold requirement is daunting, however. The current hotline system has created a triage nightmare, with 7.4 million children reported to hotlines in the most recently reported data year. Substantial screening does occur, but 3.5 million allegations annually are deemed valid enough to be sent for investigation or differential response. (HHS Child Maltreatment Report, 2016.)

A simple example shows how a hotline call can be “valid” without being “warranted”:

Assume that an active, healthy and bright 4-year-old gets frequent bumps and bruises on the playground at his preschool. Assume the 4-year old’s father is engaged in a custody battle with the mother and takes every new bruise as evidence to be used to gain an advantage in his ongoing custody battle (the sexes of the parents could be flipped here). The father (or the father’s surrogate—a therapist, doctor, or friend) calls the child abuse hotline reporting suspected abuse by the mother due to the new bruises on the child’s shoulder and legs.

There is no doubt that this abuse allegation is “valid.” If the mother intended to cause bruises to the child, the father’s allegation presents a claim that “if true, constitutes abuse or neglect.” Child abuse reporting laws started with the paradigmatic example of the “battered child” in mind (see C. H. Kempe et al., “The Battered Child Syndrome,” Journal of the American Medical Association, July 7, 1962).

Should this call therefore be investigated? I don’t think so. First, this hotline call has none of the hallmarks that turn a valid claim into prima facie true one. Its source is biased: the caller (or his surrogates; it is common for surrogates to be enlisted in the fray) seeks an outcome that is advantageous to him. Second, there is no reported history that would lead a reasonable person to believe that the mother caused the bruises, and there are many other possible causes. Playground incidents caused past bruises, which appear on common spots for accidental injuries. Third, the fear that animates need for child protection intervention into family life in the first place is not present as to this child—he is happy, healthy, and developing well.

What benefit to the child will a CPS investigation bring? If anything, the investigation will disrupt the child’s home life, adding considerably to the mother’s stress and deepening the custody battle, wasting ever more resources in a mushrooming conflict that adds the preschool and state casework teams to the parents’ personal conflict. In my opinion, this admittedly valid hotline call is not one that warrants investigation. Opening an investigation runs contrary to the child’s best interests. It’s costly to other genuinely abused children too, because the CPS worker investigating this case will have less time to build the case against the parent who is keeping his child in a cage.

Minor accidental injuries and parenting flaws that do not rise to the level of dangerous child abuse are prevalent. One of the most significant cases I worked on involved a father who, in the course of contending for custody, claimed the mother had locked her 9-year-old in a closet. He presented a perfectly valid allegation, except that the alleged closet, on quick examination, had no locks. Neglect claims against the mother, Julie Q., were deemed “founded,” however, because she was a recovering alcoholic whose sobriety claims were not believed by the caseworker and the initial reviewing administrative law judges. Four and a half years later, her case resulted in a sweeping Illinois Supreme Court determination that CPS’s neglect rule used in the finding against her (i.e., defining neglect as creating an “injurious environment”) was invalid. Further class litigation resulted in over 26,000 parents and caregivers getting similar findings removed from the child abuse register. See Julie Q. v. DCFS, 2013 IL 113783; Ashley M. v. DCFS, 2013 CH 20278 (class action); Etonia C. v. DCFS, (class action).

An especially troubling variant on this sort of hotline call occurs when the mother is fleeing domestic violence and where the hotline call is part of batterer’s attempt to maintain power and control. D. Redleaf, “Protecting Mothers Against Gender Bias” (Three Part Series), Children’s Rights Litigation, E-Newsletter, October 2011-July 2012. CPS systems have historically mishandled these cases in a “blame the victim” manner. See Nicholsen v. Scoppetta, 344 F.3d 1542nd Cir. 2003), see also S. Schechter and J. Edelson, “Effective Intervention in Domestic Violence & Child Maltreatment Cases: Guidelines for Policy and Practice,” also known as the “Greenbook,” National Counsel of Juvenile and Family Court Judges) (1999).

I’m not “insinuating” that parents and children are contacted in baseless or misplaced cases like these – I’m asserting it. Indeed, it’s the norm that children in opened CPS investigations must be seen within 24 hours of a Hotline call. Children are then interviewed by strangers, sometimes undressed, and sometimes photographed without parental consent; see Doe v. Woodard, No. 18-1066 (10th Circuit, argued September 24, 2018) (pending in the 10th Circuit on appeal from the grant of qualified immunity). Opened investigations typically stay open for 60 or 90 days, including the majority of cases that get deemed “unfounded” at the end of the investigation. In physical and sexual abuse cases, children are questioned at length about their home lives, often pulled out of class for this purpose, and then examined by state-funded doctors, a highly traumatic experience for the child. See Camreta v. Greene, 09–1454 (U.S. Supreme Court, May 26, 2011); the case was deemed moot after argument, where 43 states joined the Oregon CPS system petitioner in seeking liberal permission to interrogate children at schools without a warrant, without probable cause, without parental consent, and without exigent circumstances.

Parents report understandable anxiety when CPS caseworkers knock on their doors to investigate a child abuse or neglect allegation against them. They readily become terrified that they may lose custody of their children even when such an outcome is very unlikely. Indeed, the threat used to secure safety plans is especially potent for the already frightened parent whose sole goal is to minimize trauma to their child.

Of course, if the child in question is being beaten to a pulp, used as a prostitute, caged, or starved, any anxiety the parent or child might feel about talking to CPS is outweighed in the legitimate balance of concern for the child. Incidental trauma to the family may be justified in cases well-grounded allegations of serious harm to children (severe physical abuse, sexual abuse, and abandonment). The overwhelming number of cases aren’t these, however. Most cases CPS investigates, the data show, are not abuse or neglect at all. Of the 3.5 million children subject to investigations by CPS in 2016, 2.8 million were deemed “non-victims.”

Hotline calls may seem benign and well-intended unless and until one’s own parenting comes under the microscope. Should we be devoting precious CPS resources to investigating every valid hotline call? I don’t think so.

Unfortunately, a constant drumbeat by child advocates, fueled by celebrated child deaths “known to the system” has caused ever-expanding numbers of “see something say something” hotline calls and simultaneously has made meritless calls more difficult to screen out. The system rightly takes professional mandated reporter calls seriously. Yet 16% of calls come from unidentified and anonymous sources that bear little indicia of reliability; these calls are founded at rates averaging just 1.5%; see Dale Margolin Cecka, “Abolish Anonymous Hotline Reporting,” 64 Cath. U. L. Rev. 51 (2014). While generally better grounded, mandated professional callers can abuse the hotline too, as a recent article in the Hechinger Report/Huffington Post shows, documenting a series of cases in which teachers or school administrators made dubious calls that encouraged behaviorally challenged and disabled students to go to school elsewhere. R. Klein and C. Preston, November 17, 2018, available at

Dwyer is also mistaken in claiming that founded cases all come into court, at least until differential response deflections came along. Most founded cases are simply registered in a child abuse register and closed without anyone seeing a judge. The HHS Child Maltreatment report for 2016 shows services were provided in just slightly more than half of the founded cases, with fewer than one quarter of the founded cases resulting in foster care. The definition of “founded” cases in many states is very relaxed, and the evidentiary burden varies considerably by state. Systems also vary dramatically in the independent review mechanisms available to challenge the basis for a founded decision. Millions of findings of abuse or neglect get registered and retained without any prior neutral adjudication of guilt. See D. Redleaf, “Child Abuse Registers Abuse Due Process,” Verdict Magazine, April 2018 (available at

Findings of abuse or neglect should not make their way straight from caseworkers’ desks into state-maintained registers that operate as employment and volunteering blacklists. A qualified, neutral judge should determine guilt first. If there is probable cause to believe a child is abused or neglected by their parent or guardian, and a basis for demanding that the parent’s treatment of the child should change, then court action is appropriate. Without meeting that level of evidence, parents and children should simply be given the “right to be left alone,” since that right is the “beginning of all freedom.” (Public utilities Commission v. Pollak, 343 U.S. 451, 467 (1952)(Douglas, dissenting).

Dwyer’s discussion of Dupuy I—the phase of the case that challenged child abuse registries (Dupuy II challenged safety plans) as impairments of employment opportunity (due to their use as blacklists) is also confusing. Children as young as six could, at the time of the suit’s filing, be listed as child abuse perpetrators in registered decisions that lasted for up to 50 years. Sarah Dupuy herself was a 10-year-old who helped her mother in her day care home and was compelled to stay in another town during the investigation, under a CPS threat that the family’s day care business otherwise would close. The case did not challenge “preliminary findings”; it focused on founded guilt determinations that were placed in child abuse registers based on “practically nominal” amounts of evidence (in the language of the federal court, 141 F. Supp. 1090 (N.D. Ill. 2001).

It’s true, however, as Prof. Dwyer notes, that exoneration at the end of a child abuse investigation does not establish innocence; it merely means the state has failed to meet its burden of proof. Indeed, that’s one of the terrible consequences for parents of getting caught up in a child abuse allegation: names and reputations cannot be fully restored; corrective remedies don’t give the accused back lost time with children or restore lost careers.

It’s also virtually impossible to determine what percentage of all “founded” cases are accurate. Definitions of allegations vary by state. Systems to challenge “founded” decisions are rudimentary and difficult to navigate in many states (Illinois, post Dupuy, hears 20 times more register appeals than California does). Overturning rates vary dramatically from state to state as well. Advocates like me would welcome a major neutral study of how many millions of parents are wrongly listed in child abuse registers, but faulting me for a lack of data that isn’t maintained is a bit unfair. Setting a proper balance between Type I (false positive) and Type II (false negative) errors will not occur, however, if false positives are treated as harmless and discounted, and if all attempts to fix the over-intrusion problem are dead on arrival.

Understanding the potential negative consequences for families who get investigated and registered as perpetrators of abuse or neglect in error leads me to view community service referrals through differential response programs much more positively than Prof. Dwyer does. The record on differential response is less bleak than Prof. Dwyer suggests (as Richard Wexler’s comment notes). The research that Professors Dwyer and Bartholet cite is both equivocal and non-standardized, with differences in reporting of subsequent maltreatment relatively small and specific to state systems that may have improperly referred serious abuse cases to differential response in the first place.

Turning to the remaining questions Dwyer presses me to answer:

(1) There is nothing inconsistent between my claims that CPS over-investigates meritless claims while genuine abuse or neglect is underreported, as found in National Incidence Studies. Our shared goal should be accuracy and fair targeting.

(2) Safety plan separations continue in many states including Illinois, because despite the settlement of 5 post-Hernandez lawsuits, implementation of agreements has not been complete. There have been recent news reports of these separations in Delaware, Arizona and Texas.

Prof. Dwyer challenges me to expand on a parent/child/state jurisprudence. While neither of us has space in this forum to satisfy this challenge, my starting point would be different from Prof. Dwyer’s. Parent and child familial rights are properly treated as fundamental. Professor Dwyer wishes to elevate a child’s right to safety, judged by fallible state officials, as overriding parent’s (and child’s own familial rights). How else am I to read his proposed state “fiduciary duty” to create for children an appropriate home life? Much as I am dismayed by the current state of CPS, I’m even more troubled by a system that would generally empower the state to decide when any children or parents deserve each other, with children given a say on an equal footing with parents no matter their age.

I view families as groups of human beings who should presumptively be stuck with each other, for better and sometimes for worse. When it is for worse, the state’s powers still need to be carefully circumscribed, with strict scrutiny and higher burdens of proof than our CPS systems currently endorse. Empowering child representatives to speak for young children who cannot speak for themselves is a wonderful full employment system for lawyers and well-meaning volunteers, but does little to advance a fair adjudicative process. See M. Guggenheim, “The Right to be Represented but Not Heard,” 59 N.Y.U.L. Rev. 76 (1984). 

As a family defense lawyer, I take due process seriously. The state should have no inherent powers to intervene in family life. The state should be required to have at least probable cause to believe the parent has seriously mistreated or immediately endangers a child before it begins to limit the parent’s rights to decide what’s best for that child. Limits should be judicially enforced, not issued by caseworker fiats. I don’t advocate this scheme solely because it is a best for parents: it’s also best for children and for our society.


When I was invited to write my opening essay, I chose one of the themes in my new book (They Took the Kids Last Night: How Child Protection Systems Put Children at Risk)—unwarranted safety plans issued without due process. I worried, however, that there wouldn’t be much to debate. Prof. Dwyer’s searching questions and critiques surprised me.

While Professor Dwyer and I differ substantially in our emphases and assumptions, we might agree substantially on what a reasonable lawful CPS system looks like. Do we both think well-grounded cases of abuse or neglect should come to court? I think so. Do we both worry about out-of-court arrangements that deflect cases into informal alternatives—yes, I focus on informal alternatives that needlessly separate families, while Dwyer is more concerned with deflections that keep children in unsafe homes. Do we both think investigations should happen in every case—maybe so, if the definitions of maltreatment were clearer to everyone and prima facie evidence were required first. Should parents be given options to cooperate or not, without being forced into court: yes, assuming they are not being baselessly threatened and if they, like major corporations negotiating in the shadow of the law, have meaningful access to counsel.

Do all of our agreed upon remedies require a better trained CPS workforce? Definitely, though training won’t solve confusions about what we expect CPS to do. Due process for the accused would answer many concerns.. Shadow processes need to be out in the sunlight. Maybe Professor Dwyer and I do not disagree so deeply after all.

Also from this issue

Lead Essay

  • Diane Redleaf describes how Child Protective Services operates, using one particularly troubling case as a focus. Yet millions of children are named as suspected victims of abuse every year, and many of these end up being removed from their families, whether temporarily or permanently. Redleaf describes how parents may be subject to strong tactics to secure removal—and how these tactics can easily affect the innocent.

Response Essays

  • James G. Dwyer says that Child Protective Services may already err in favor of the parents. Child separations are never an easy call, but the incentives at hand, Dwyer argues, are likely already arranged such that in general children are not removed from a home more often than they should be. Indeed it’s quite possible that the opposite is true, Dwyer says.

  • Walter Olson finds that the problems with Child Protective Services are of a kind with problems that affect many other agencies of government. He draws readers to consider these similarities and proposes several ideas that aim at reform. On many of his points, transparency and accountability emerge as key considerations. Bureaucrats seldom want these, but bureaucracy quickly grows abusive without them.