A syndicated “data” piece says family courts fail to protect domestic violence victims because of implementation, not law. The studies behind it say something larger.
By Michael Phillips | Family & State
On September 21, radio stations and news sites carried a piece headlined “Domestic violence in family courts: What the data reveals about protection gaps.” It was written by a Texas custody law firm and distributed through Stacker. Its conclusion is that the gap between the prevalence of intimate partner violence and the protection courts deliver is “not primarily a gap in legal authority.” It is a gap in implementation, resources, and consistency.
That conclusion is correct. But the studies underneath it describe something the piece never says: the failures these studies describe are structural, so they can land on accused parents as well as on victims. Inconsistency doesn’t pick sides.
The piece itself describes these gaps as structural. Follow that premise one step further and an uncomfortable question appears: what happens when an inconsistent system gets it wrong in the other direction?
Old studies, regional samples
The piece reads as a current snapshot. The research is not.
The custody-evaluator study draws on 69 New York City cases litigated and resolved between 1997 and 2007; the research project itself ran through 2009. The protective-order study that produced the “40% cited judicial bias” figure covers five Kentucky jurisdictions and was published in 2010. The Iowa violation-rate study tracked 768 defendants during six months of 2009.
These are serious studies, funded and published by federal and state agencies. They are also regional, narrowly sampled, and about 15 years old (some of the underlying data is older), and they are being presented as the national picture. Readers deserve to know that.
“Inconsistency doesn’t pick sides.”
The evaluator problem
When a custody case includes domestic violence allegations, the court often appoints a psychologist, psychiatrist, or social worker to evaluate the family. The National Institute of Justice funded a study of how those evaluators work, based on 69 cases from four New York legal services organizations that represent domestic violence victims.
Its findings are striking. Evaluators held “a wide range of beliefs” about domestic violence and children’s best interests, and those beliefs shaped their conclusions and the court outcomes that followed. According to a summary of the report, the strongest predictor of whether a parenting plan was safe was the evaluator’s own assessment of ongoing risk, and the evaluators’ recommended plans did not differ significantly from the final settlements or court orders. In practice, the evaluator’s recommendation generally was the outcome.
The researchers’ own conclusion: a family’s fate should not depend on which evaluator is appointed.
That sentence applies in every direction. A victim paired with an evaluator who doesn’t understand coercive control is failed. A parent facing an unproven allegation, paired with an evaluator who treats allegation as finding, is failed too. The mechanism is the same.
The study cannot tell us how often the second scenario occurs: its sample came only from victim-side legal services, and it wasn’t designed to measure respondent-side error. Its remedy, though, covers both. The researchers recommended screening and training evaluators on domestic violence, and also recommended that courts conduct their own fact-finding rather than relying on evaluators to investigate. A judge hearing evidence and making findings creates an accountable judicial record. An evaluator’s impression is not a substitute for that process.
“A family’s fate should not depend on which evaluator is appointed.”
“Who you know”
The 40% figure deserves a closer look. In the Kentucky study, 188 key participants, including judges, police, prosecutors, defense attorneys, court clerks, and victim-services workers, were asked about barriers to obtaining a protective order. Forty percent mentioned “judicial bias.”
The study’s own description of bias is revealing. It includes a judge’s personal political connections to the families involved, and how a judge treats a woman who has filed more than once. Rural respondents described a “who you know” system.
Notice what that describes. It is arbitrariness: outcomes that turn on relationships and reputation instead of evidence. It was documented as a barrier for women seeking orders. But a system where results depend on connections is a system where anyone can lose. The study documented this only from the petitioner’s side and did not measure its effect on respondents, but a system run on connections is hard to trust in either direction.
“A system where results depend on connections is a system where anyone can lose.”
The study also carries a limitation worth stating plainly. Its survey of women covered only those who obtained orders, so it cannot say what happened to those who were denied.
The same research also found that protective orders generally improved safety and reduced fear. Half of the women had no violations in the following six months, and abuse fell significantly among those who did. The finding was not that protective orders generally fail; it was that access, enforcement, and outcomes varied substantially.
What a “16% violation rate” can and can’t tell you
The Skillern piece cites an Iowa study finding a 16% violation rate for protective and no-contact orders, and treats it as evidence that enforcement depends on victims reporting breaches.
The figure is real. Iowa researchers found it among 768 defendants during the first half of 2009, along with a 15.6% conviction rate for subsequent domestic abuse offenses and variation in conviction rates between judicial districts. But one state’s six-month study is not a national constant. A meta-analysis of 25 studies found that violation rates were higher when measured from victim reports than from police reports, and that researchers have no agreed way to measure order effectiveness at all.
The abstract of Iowa’s own report acknowledges the problem: violation rates “have varied widely from study to study.” A number that varies that much depends on who is counting and how.
“A number that varies that much depends on who is counting and how.”
What nobody measured
Here is the largest gap in the three studies examined here. As far as their published summaries show, none of these three studies set out to measure how often allegations turn out to be unfounded, or how often an ex parte order or a flawed evaluation harms a parent who did nothing wrong. They weren’t designed to. The evaluator and Kentucky studies drew on victim-side samples, and the Iowa study tracked defendants but asked how often orders were violated, not whether they were warranted.
That means the data cannot support the claim that courts fail victims more than they fail the accused, or the reverse. It supports something narrower and more useful: outcomes in these cases vary by evaluator, by county, by judge, and by who is doing the counting. A system capable of producing unpredictable results can produce errors in either direction, and without error-tracking nobody can say how many.
Courts weighing abuse allegations face two kinds of error, and they are not equal. Failing to protect a genuine victim can leave a parent and child exposed to continuing violence. Restricting a parent on an allegation that cannot be substantiated also carries serious consequences. These studies do not quantify or compare the two. “No side” does not mean equal frequency or equal severity. It means a decision system has to care about error in both directions.
“‘No side’ does not mean equal frequency or equal severity.”
What would close the gap
“A decision system has to care about error in both directions.”
The first two reforms below follow directly from the research; the last two follow from the gap in the data:
- Judicial fact-finding on abuse allegations instead of delegation to evaluators, as the NIJ researchers recommended.
- Uniform evaluator standards, covering screening, training, and documentation of how conclusions were reached.
- Consistent, public data on protective order grants, denials, violations, and outcomes, tracked the same way across counties.
- Prompt adversarial review after emergency ex parte orders, so a temporary one-sided proceeding is followed quickly by a chance for both parties to present evidence.
None of that weakens protection for genuine victims. It makes protection something a court can defend.
The Skillern piece asks readers to see a gap between authority and outcomes. They’re right. The gap is real. These studies tell us a great deal about where the system can fail victims. They do not establish that institutional inconsistency can produce error in only one direction.
Disclosure: The author is a parent whose own custody case involved protective orders and has discussed that experience publicly. This piece relies only on the published research cited above.
Sources
- NIJ, Custody Evaluations When There Are Allegations of Domestic Violence: Practices, Beliefs and Recommendations of Professional Evaluators (O’Sullivan et al.), NCJ 234465
- NIJ Journal No. 266, “Perspectives on Civil Protective Orders in Domestic Violence Cases: The Rural and Urban Divide” (June 2010), NCJ 230410
- Iowa Department of Human Rights, The Effectiveness of Domestic Abuse Protective Orders & Court Practices in Sentencing Violators (July 2011), NCJ 238662
- Trauma, Violence, & Abuse, systematic review and meta-analysis of protection-order effectiveness in reducing recidivism
- KNSI / Stacker syndication of the Skillern Firm piece (Sept. 21, 2026)
Keep Family & State Free
Family & State exists to support parents navigating separation, custody, and systems that are often confusing, isolating, or overwhelming. This work is grounded in lived experience, careful research, and respect for the real stakes families face.
If this article helped you feel less alone, better informed, or more grounded, reader support helps keep these resources free and available to others who need them.
Need help reviewing or organizing court or formal documents?
Family & State offers non-legal document review and organization for people representing themselves. This includes clarity, structure, neutral tone, and timeline organization — not legal advice or representation.
Have a story, experience, or resource to share?
Submissions are reviewed with care and discretion. We respect privacy and handle sensitive information responsibly.
Discover more from Family & State
Subscribe to get the latest posts sent to your email.
