What the research can support
What the evidence says—and what it cannot prove.
A case can be marked closed and still not be truly resolved. Court statistics can reveal workload and delay. They cannot, by themselves, tell us whether the result was safe, fair, affordable, or workable for the family.
Numbers show patterns, not individual truth. On this page, every statistic stays attached to its place, time period, definition, and limits. When a study finds two things are connected, we say “associated”—not “caused”—unless the research can support causation. Research terms explained →
How to read this page
Pressure signals are not verdicts.
A clearance rate below 100% means fewer cases were closed than were added or reactivated during that period. It does not tell us why. A long case may reflect necessary safety work. A short case may reflect efficient resolution.
Or someone giving up.
That is why FJP pairs workload measures with safety, fairness, cost, and what happens after an order.
Verified directly supported by the cited record. Inference our clearly marked interpretation or idea to test. Limit something the record cannot tell us, a missing fact, or another reasonable explanation.
Administrative data · Illinois
Open family caseload grew in 2024.
See the statewide shift
Choose a year. Each bar uses the same 90,000-case scale.
In 2024, the ending open caseload was 8,974 cases higher than in 2023—a rise of 11.8%.
| Measure | 2023 | 2024 | What can be said |
|---|---|---|---|
| New family cases filed | 59,380 | 58,679 | New filings fell 1.2%. |
| Cases closed | 50,002 | 49,507 | Closures fell 1.0%. |
| Clearance rate | 83.5% | 83.6% | Closures remained below incoming and reinstated volume. |
| Ending open caseload | 76,272 | 85,246 | Open cases increased by 8,974 (11.8%). |
Verified These are statewide “family” category counts in the Illinois Courts 2024 Annual Report. The category definition changed before 2022, so this project avoids a longer historical series without harmonized definitions. Source S01
Limit Aggregate filings do not distinguish protective complexity from remediable process delay, and “open” does not mean continuously contested.
A 1998 U.S. Government Accountability Office review reported that an independent committee examining a prominent Cook County child-death case identified the collective failure to provide crucial information as a major systemic failing. Case information was kept in hard-copy files; heavy caseloads left judges and attorneys unfamiliar with the record; and the last judge assigned had no prior family-court experience. The report said an automated information system might have made a significant difference. Source S40
Boundary This is a historical child-protection example, not evidence of current Illinois private-custody performance. Its continuing relevance is the failure mode: fragmented information plus discontinuity can defeat safeguards that formally exist.
Multi-court study · United States
Reopening and self-representation are structural design questions.
The NCSC/IAALS Landscape study examined 147,436 completed divorce and other domestic-relations cases in 11 large urban courts—roughly 8% of the national total at the time. It reported that most people did not hire attorneys, contested and uncontested cases took about the same time, and one in four cases later returned to court. Returning is not automatically failure; a change may be necessary. The researchers also found court data too inconsistent to answer many basic questions. Source S02
| Finding | Implication for a pilot | Not established |
|---|---|---|
| About 25% reopened | Measure durability and postjudgment hearings, not just first disposition. | That all reopenings are failures; modification may be appropriate. |
| Contested and uncontested cases took about the same time | Test whether path assignment reflects actual complexity. | That contest alone caused delay. |
| Data were incomplete/inconsistent | Use a defined minimum dataset and publish a data dictionary. | That missing data imply misconduct. |
Current domestic snapshots
The American data are large—and still frustratingly incomplete.
| Public record | Reported measure | Boundary |
|---|---|---|
| California, FY 2023–24 | 108,403 marital filings and 210,326 other family-law filings. | Volume is not delay or quality. “Other” includes matters such as parentage and child support. |
| Washington Superior Court, 2025 | 24,113 domestic filings in the published statewide table. | King County is shown as zero, so the table should not be treated as complete statewide demand. |
| California, Apr. 2023–Mar. 2026 | Of 4,214,365 hearings across family law, probate, and major civil matters, 71.3% had no word-for-word transcript or recording. | These court categories are combined. No word-for-word record does not mean there was no written minute order or other summary. |
Source S34 Source S33 Source S35
The missing number is sometimes the finding.
When a major county appears as zero, categories change, or millions of consequential hearings lack a verbatim record, the first reform need is not a prediction model. It is a trustworthy data foundation.
International comparator · England and Wales
A live administrative series shows the value of aging data.
For January–March 2026, private-law children cases averaged 35 weeks to final order. At quarter-end, 40,971 cases were open and 1,692 (4%) had been open at least 100 weeks. New-case starts rose 16% year over year while disposals fell 17%. In 49% of private-law disposals, neither party had recorded legal representation; both parties were represented in 13%. Source S03
These Ministry of Justice figures describe England and Wales, not the United States. They are included to demonstrate a current, transparent aging series—not as an estimate of U.S. delay.
Cost and access
When truth becomes unaffordable, procedure can change substance.
The Legal Services Corporation’s 2022 Justice Gap survey found that low-income Americans received inadequate or no legal help for 92% of their substantial civil legal problems; among those who did not seek help, 46% cited cost and 53% doubted they could find an affordable lawyer. The survey covered civil legal needs broadly, not family court alone. Source S04
An unpublished Illinois appellate order records about $165,000 in one spouse’s attorney time over three years. This is a documented illustration—not a typical-cost estimate, not precedential authority, and not evidence that all fees were avoidable. Source S05
Two U.S. Supreme Court cases show how money and procedure can intersect with fundamental family interests: Turner v. Rogers addressed incarceration after a child-support contempt hearing without adequate ability-to-pay safeguards; M.L.B. v. S.L.J. addressed an appeal blocked by record-preparation fees after termination of parental rights.
Children and families
Conflict is associated with harm; “litigation caused it” requires more proof.
A 2020 meta-analysis combined 115 samples from divorced families—24,854 people in all. Conflict between parents, parenting behavior, and children’s adjustment were connected, usually by small amounts. Some of the connection appeared to operate through changes in parenting. Much of the research observed families rather than randomly assigning a process, so it cannot prove that court proceedings caused the outcomes. Source S06
A 12-year follow-up of 109 parents originally randomly assigned to mediation or adversarial settlement found early decreases in reported coparenting conflict after mediation and increases after the adversarial path. The study is small, old, and context-specific; mediation is not safe or appropriate in every case. Source S07
A separate randomized study of families reporting high intimate-partner violence found no significant one-year group differences among shuttle mediation, videoconference mediation, and litigation on continuing violence/conflict, satisfaction, parenting, or parent/child functioning. That null result is an important warning against one-size-fits-all claims. Source S08
The CDC’s 25-state analysis of adverse childhood experiences found strong links between accumulated childhood adversity and later health problems. For example, adults reporting four or more adverse experiences had substantially higher odds of depression. The information was collected at one point in time, depended on people’s own reports, and was not specific to family court. It supports reducing avoidable childhood adversity—not a claim that litigation caused those outcomes. Source S09
Existing law and standards
FJP builds on safeguards that already exist.
- Illinois evaluation law: court-appointed evaluation reports must include procedures, data, results, conclusions, recommendations, and limitations; the court defines scope. S16
- Illinois fee transparency: certain child representatives and attorneys for children must submit detailed invoices at least every 90 days; courts approve reasonable and necessary fees. S17
- Professional standards: APA custody-evaluation guidelines emphasize competence, multiple data sources, relevance, and attention to intimate-partner violence, child maltreatment, and substance use. S18
- Performance measurement: NCSC family-justice measures pair timeliness and pending-age data with procedural fairness, postjudgment activity, and cost from both court and party perspectives. S19
FJP’s contribution is to connect those principles into a common, testable operating model—while treating jurisdiction-specific law as controlling.
Evidence gaps
What must be learned before scale.
- Reliable U.S. national estimates of family-litigation duration, private cost, and re-litigation using harmonized definitions.
- Independent, prospective research that separates pre-existing family risk from harm added or reduced by process design.
- Validated screening and triage performance across culture, disability, language, income, and gender.
- False-positive and false-negative rates for restrictions and safety interventions.
- Whether evidence-ledger and scope controls improve accuracy without adding delay or bureaucracy.
- How to measure child stability without invasive or coercive follow-up.
These gaps are not reasons to do nothing. They are reasons to test one small part at a time, write down the questions and measurements before seeing the results, protect participants, and publish disappointing findings as well as positive ones.