From Incidents to Trajectories: What Robert Kaiser’s Research Could Teach the Family Court About Coercive Control Evidence

Research acknowledgement

This article responds to and builds upon Robert Kaiser’s article Reconstructing Behavioral Trajectories in Intimate Partner Violence: What Could This Add to Police Investigations?, published on 2 October 2026.

Kaiser is the founder of the Women’s Safety Institute and developer of Escalation Pattern Analysis (EPA), a trajectory-based framework concerned with how abusive behaviour changes over time.

The application to Family Court evidence discussed below is JSH Law’s analysis. It should not be read as Robert Kaiser’s endorsement of JSH Law, or as suggesting that EPA itself is a Family Court evidential or legal test.

Why this matters for Family Court evidence in 2026

The direction of family procedure is already moving towards pattern analysis.

Practice Direction 12J now expressly provides that where a schedule or table of domestic-abuse allegations is used, it may contain examples of an alleged pattern of coercive and controlling behaviour so that the judge can consider the alleged abuse holistically rather than incident by incident.

That makes the central question behind Kaiser’s work particularly timely: what becomes visible when we reconstruct behaviour over time?

Five things to understand first

1. Domestic abuse evidence is often fragmented.
Police, schools, medical professionals, Cafcass, social workers, digital platforms and the parties themselves may each hold only part of the history.

2. Chronology can reveal something individual incidents cannot.
Frequency, changing severity, increasingly intrusive behaviour and responses to previous boundaries may become visible only when events are reconstructed over time.

3. Pattern does not remove the need for proof.
Placing events together cannot turn an allegation into a fact. Source, status and attribution still matter.

4. EPA is not a Family Court test.
It is an emerging analytical framework developed by Robert Kaiser. It must not be presented as though the Family Court has adopted it as a validated risk-assessment instrument.

5. The practical lesson is nevertheless powerful.
A litigant in person should help the court understand not only what happened, but where appropriate how the relevant behaviour developed over time.

What is Robert Kaiser arguing?

Kaiser’s starting point is deceptively simple.

In a serious intimate-partner-violence investigation, the problem is frequently not the complete absence of information.

The information already exists.

But it exists in different places.

A police officer records one incident.

A doctor records an injury.

A neighbour remembers an argument.

Messages reveal monitoring or threats.

A safeguarding agency has information about another intervention.

Children may have witnessed or experienced aspects of the behaviour.

Each organisation records information for its own purpose.

Nobody necessarily reconstructs the entire history.

Kaiser asks what happens if we do.

Instead of asking only:

“What happened in this incident?”

we also ask:

“How has the behaviour changed over time?”

That is a materially different investigative question.

What is Escalation Pattern Analysis?

Kaiser developed Escalation Pattern Analysis (EPA) as a trajectory-based approach to understanding escalating intimate partner violence.

The current framework examines behavioural change through three interacting dimensions:

Kaiser deliberately moved EPA away from numerical scoring.

He describes it as a non-scoring, non-categorical and non-predictive framework intended to support professional judgment rather than replace it.

That limitation matters.

It is one of the reasons I find the work particularly interesting.

This is not another attempt to create a magic number which claims to tell us whether somebody is “high risk”.

It asks us to examine direction of travel.

The Family Court has the same fragmentation problem

Although Kaiser’s latest article is directed principally towards police investigations and later prosecution, the fragmentation he describes will be instantly recognisable to anyone working with complex private children proceedings.

A Family Court case may contain information spread across:

• C100 and C1A forms;
• historic court orders;
• previous findings or admissions;
• police disclosure;
• crime or incident numbers;
• GP records;
• hospital records;
• school safeguarding records;
• local-authority records;
• Cafcass or Cafcass Cymru assessments;
• Child Impact Reports;
• domestic-abuse-service records;
• WhatsApp messages;
• emails;
• telephone records;
• smart-device or location evidence;
• financial records;
• witness statements;
• previous position statements; and
• the recollections of family members or other witnesses.

Each source may contain one fragment.

And the documents may have been created months or years apart.

This creates an obvious danger.

A pattern can disappear inside the paperwork.

Why Kaiser’s work connects so strongly with current Family Court procedure

This is not merely an interesting academic parallel.

Current Family Court procedure expressly recognises the problem of incident-by-incident analysis.

Practice Direction 12J defines coercive and controlling behaviour in terms which expressly recognise patterns of acts.

And following the January 2026 amendments, paragraph 19 specifically provides that a schedule or table may set out:

specific events and examples of any pattern of coercive and controlling behaviour so the judge can consider the alleged abuse holistically rather than focusing upon individual incidents alone.

That is a significant procedural development.

It means a document which simply says:

may still fail to communicate what the case is really about.

The court may also need to understand:

• what preceded those events;
• whether they became closer together;
• whether boundaries had previously been communicated;
• whether a court order or police warning intervened;
• what happened after that intervention;
• whether behaviour moved into new domains;
• how the child was affected; and
• what remains allegation rather than established fact.

That is trajectory thinking.

It is not the same as proving the allegation.

But it may help the court see the question it actually needs to determine.

1. Frequency: is the behaviour becoming more compressed?

Frequency is more sophisticated than simply counting incidents.

The timing may matter.

Imagine alleged unwanted contact occurring:

• once every few months;
• then once a month;
• then weekly;
• then several times within one week.

The raw behaviour may appear similar.

But the temporal pattern has changed.

Kaiser’s work describes this in terms of increasing frequency and temporal compression.

For Family Court evidence, that may justify asking:

• Are events becoming closer together?
• Is the same conduct recurring?
• Does it occur around particular triggers?
• Does it increase around handovers, hearings, new relationships, applications or attempts to establish boundaries?

Those questions must still be answered by evidence rather than assumption.

2. Intensity: has the behaviour changed in character?

Escalation does not necessarily mean that the same act simply happens more often.

It may change in quality.

For example, an alleged sequence might move from:

repeated messages → monitoring → unwanted attendance → threats → breach of an imposed boundary

Or economic behaviour might develop from:

questioning spending → restricting access → withholding resources → using financial dependence after separation

Again, a chronology does not establish why those events occurred.

But it allows the court to examine whether the nature of the alleged behaviour changed.

That is often lost when every allegation occupies one disconnected row in a schedule.

3. Inhibitory decline: what happens after a boundary is imposed?

This is the aspect of Kaiser’s framework which I think deserves particularly careful consideration.

He describes inhibitory decline as the weakening of restraints which previously appeared to limit behaviour.

In his recent article, he reduces the investigative question to a powerful line:

“What used to stop him, and does it still?”

For Family Court work, I would phrase the evidential question slightly more neutrally:

What boundary, intervention or consequence occurred — and what observable behaviour followed?

Examples might include:

• a police warning;
• a non-molestation order;
• a child arrangements order;
• a solicitor’s written boundary;
• contact moving to a supervised setting;
• a school asking a parent not to attend;
• a domestic-abuse intervention;
• separation;
• a family member intervening; or
• the other parent changing communication arrangements.

The critical point is that a litigant should describe the observable response.

They should not speculate about the other person’s internal psychology.

The first describes evidence.

The second potentially ventures into expert risk opinion which the litigant may not be qualified to give.

What might a behavioural trajectory look like in a Family Court case?

Consider a hypothetical private children case.

A parent alleges post-separation coercive control.

If the evidence is presented incident by incident, it might look like this:

Those events may or may not establish domestic abuse.

But trajectory analysis asks additional questions:

• Are incidents occurring closer together?
• Has the behaviour moved across communication, physical attendance and third-party channels?
• What happened after the police intervention?
• Did the previous communication boundary remain effective?
• Is there reliable evidence identifying who carried out each act?
• How, if at all, was the child affected?
• What issue does the Family Court actually need to decide?

Those are better questions than simply counting five allegations.

Do you have years of domestic-abuse evidence spread across different documents?

This is one of the areas where careful evidence organisation can materially change how understandable a case becomes.

JSH Law can provide defined-scope support with:

• domestic-abuse chronologies;
• behavioural-pattern timelines;
• evidence and source audits;
• police / medical / school / Cafcass evidence mapping;
• schedules of allegations and responses;
• witness-statement preparation support;
• C1A preparation support;
• Cafcass or Child Impact Report responses;
• position statements;
• appeal evidence analysis; and
• hearing preparation.

The aim is not to make allegations sound more serious.

It is to make the evidence easier to trace, test and understand.

Get help organising your evidence

From chronology to trajectory: a more useful evidence map

A conventional chronology usually answers:

What happened, and when?

A trajectory-informed chronology can ask more.

For example:

I would not call this an “EPA assessment”.

It is better understood as a Family Court evidence architecture informed by trajectory thinking.

That distinction protects both Kaiser’s work and the integrity of the court evidence.

Where this intersects with the JSH Law Six-Question Check

Kaiser’s trajectory approach and the JSH Law Six-Question Check solve different but complementary problems.

Trajectory analysis asks:

How is behaviour changing over time?

The Six-Question Check asks:

Together, those questions resist two opposite errors:

• breaking a pattern into meaningless fragments; and
• constructing a pattern which the underlying evidence does not support.

The limits: what trajectory analysis cannot prove

This is crucial.

A chronology can reveal relationships between events.

It cannot automatically prove causation, attribution or intent.

If two incidents occur closer together, that does not by itself prove escalation.

If a person acts differently after a court order, that does not automatically establish why.

If several agencies record concerns, repetition across records does not necessarily mean there are several independent sources — they may all derive from the same original allegation.

And if a pattern looks obvious to one party, the court may reach a different conclusion after hearing both sides.

This is why trajectory work must remain evidence-led.

Do not turn an emerging framework into unauthorised expert evidence

There is another important legal boundary.

EPA is an emerging research framework.

It is not a legal test adopted by the Family Court.

A litigant should therefore be extremely cautious about writing:

“EPA proves that the respondent is escalating towards lethal violence.”

That would go far beyond organising evidence.

It potentially moves into expert risk opinion.

Expert evidence in children proceedings is controlled by Part 25 of the Family Procedure Rules and section 13 of the Children and Families Act 2014.

The general principle is that expert evidence must be necessary to assist the court, and permission is required before it is put before the court.

Since July 2026, additional rules also govern the use of regulated experts in relevant children proceedings.

So what can a litigant properly do?

They can organise their own factual evidence.

They can identify chronology.

They can draw the court’s attention to repetition or change which is apparent from the evidence.

They can explain why particular events matter to their case.

They can ask the court to consider an alleged pattern as PD12J requires.

What they should not do is present themselves as a forensic psychologist, lethality expert or validated risk assessor.

What could trajectory thinking mean for Cafcass and Child Impact Reports?

The same issue arises with professional family-court assessments.

A report can list incidents accurately and still fail to capture change over time.

When reading a Cafcass, Cafcass Cymru or Child Impact Report, it may therefore be useful to ask:

• Does the report distinguish allegation from finding?
• Has it identified the relevant history?
• Are current events assessed against earlier behaviour?
• Has post-separation behaviour been considered?
• Is behaviour treated as isolated “parental conflict” where a pattern of control is alleged?
• Has the impact on the child been identified?
• Does the report consider the effect of previous interventions?
• Are evidential gaps identified or silently filled by inference?

This does not mean a litigant should reject a report simply because it does not adopt Kaiser’s framework.

It means a report can be tested for whether it has understood the history, status and direction of the evidence.

Trajectory analysis can also reveal what is missing

Kaiser makes another valuable point.

Once a chronology is reconstructed, the gaps themselves may become informative.

Not because a gap proves that something happened.

It does not.

But because it may identify a legitimate evidential question.

For example:

• Was there an earlier police attendance?
• Is there a GP presentation corresponding with an alleged event?
• Did a school record a disclosure?
• Are there previous court findings or admissions?
• Was there an earlier safeguarding referral?
• Does the other party rely on an intervention or course which needs proper evidence?
• Is an apparently new incident actually part of a previously documented history?

PD12J itself recognises that relevant third-party material may include records from police, health services and domestic-abuse support services.

But this must not become a fishing expedition.

Disclosure should be relevant, proportionate and connected to an issue the court actually needs to determine.

Technology makes trajectory reconstruction both easier and harder

Modern domestic-abuse cases can generate enormous digital histories.

Messages.

Emails.

Location alerts.

Account-access logs.

Doorbell footage.

Photographs.

Voice notes.

School apps.

Bank records.

The problem is no longer simply preserving information.

It is making sense of it.

This is particularly relevant to technology-facilitated abuse, where apparently minor digital events may acquire different significance when viewed chronologically.

But the same warning applies:

the pattern must emerge from the evidence rather than being imposed upon it.

Could AI help reconstruct a behavioural trajectory?

This is where Kaiser’s work intersects directly with legal technology.

A well-designed AI workflow could potentially help a human reviewer:

• extract dates from large document sets;
• build a provisional chronology;
• identify repeated behaviours;
• calculate intervals between events;
• group documents by source;
• identify possible duplicate accounts of the same event;
• locate references to police warnings or court orders;
• compare behaviour before and after an intervention; and
• flag evidential gaps for human review.

That could be extremely useful.

But it is also precisely where AI can become dangerous.

An AI system can see patterns which are not really there.

It can mistake repetition of one allegation across four documents for four independent events.

It can infer identity from incomplete metadata.

It can turn sequence into causation.

It can convert “after” into “because of”.

And because the output sounds analytical, the error may be difficult to spot.

The safest model is therefore:

AI-assisted extraction → human verification → source checking → legal relevance → court-ready evidence

Not:

Upload bundle → ask AI who is dangerous → file the answer.

I think my case contains a pattern of coercive control. What should I actually do?

Re C makes this conversation especially timely

The Court of Appeal’s recent decision in Re C (Child Arrangements Orders; Domestic Abuse; Child Focused Court) [2026] EWCA Civ 1216 reinforces the centrality of pattern-based thinking.

Sir Stephen Cobb emphasised that controlling and coercive behaviour requires appreciation of a pattern or series of acts whose impact must be considered cumulatively rather than rarely in isolation.

The judgment also confirms that the Child Focused Court model does not dilute PD12J.

So there is an interesting convergence here.

Family procedure is moving explicitly towards holistic analysis.

Kaiser’s work asks practitioners to reconstruct how behaviour moves through time.

Both challenge the same evidential failure:

looking so closely at individual incidents that the pattern disappears.

Read JSH Law’s analysis of Re C [2026] EWCA Civ 1216.

What I think Kaiser’s work adds — and what still needs testing

The strongest feature of Kaiser’s approach is not a claim that EPA can predict the future.

In fact, its current formulation expressly resists that.

Its strength is that it forces professionals to ask a different class of question.

Not simply:

How serious is this incident?

But:

What direction is the behaviour moving in?

That is potentially valuable because static assessment can miss movement.

A case may contain no dramatic new category of behaviour while still changing materially through:

• greater frequency;
• temporal compression;
• increasing intrusion;
• movement across different forms of control;
• breach of boundaries; or
• reduced response to previous interventions.

But an emerging framework also needs scrutiny.

Questions which matter include:

• How consistently can different practitioners identify the same trajectory?
• How should contradictory or incomplete records be handled?
• How do we protect against confirmation bias?
• How do we distinguish genuine escalation from increased reporting or increased evidence availability?
• How should apparent de-escalation be interpreted?
• How do different abuse types interact?
• What is the framework’s operational reliability?
• How should survivor interpretation and professional judgment interact?

Kaiser himself recognises that further operational testing is necessary.

That is important.

Good safeguarding innovation should be capable of being questioned without being dismissed — and used cautiously without being prematurely treated as settled science.

The chronology should not merely tell us what happened. It should help us see what changed.

This is what I take from Kaiser’s work.

Domestic-abuse cases generate fragments.

The police see one fragment.

The GP sees another.

The school sees another.

Cafcass sees another.

The Family Court receives another.

The survivor has lived through all of them.

The challenge is to bring those fragments together without distorting them.

Not every sequence proves a pattern.

Not every pattern proves escalation.

Not every concerning trajectory permits a prediction about what somebody will do next.

But neither should the justice system repeatedly examine one incident at a time and then be surprised when the whole is greater than its parts.

Preserve the source. Establish the status. Reconstruct the chronology. Examine the change. Test the pattern. Show the welfare impact.

For Family Court evidence, that may be the practical bridge between an allegation which feels overwhelming and a case which can actually be understood.

Need help reconstructing a complex domestic-abuse evidence history?

JSH Law provides defined-scope support for litigants in person in private children proceedings in England and Wales.

Support can include:

• domestic-abuse and post-separation abuse chronologies;
• evidence-source registers;
• behavioural-pattern timelines;
• contradiction and evidential-gap analysis;
• schedules of allegations and responses;
• C1A preparation support;
• witness-statement preparation support;
• Cafcass and Child Impact Report analysis;
• position statements;
• appeal evidence analysis;
• bundle preparation support; and
• hearing preparation and McKenzie Friend support where appropriate.

The aim is not to produce more paperwork. It is to make the evidence easier for the court to understand and test.

Book a consultation

Related JSH Law analysis

• Technology-Facilitated Abuse in the Family Court: GPS Tracking, Smart Devices & Digital Evidence
• Child Focused Courts Must Still Be Fair: Re C [2026] EWCA Civ 1216
• Domestic Abuse in the Family Court: What the IDAS Family Court Report 2026 Reveals
• The JSH Law Six-Question Check

Research, legal and official sources

• Robert Kaiser — Reconstructing Behavioral Trajectories in Intimate Partner Violence: What Could This Add to Police Investigations?
• The Women’s Safety Institute — Robert Kaiser and Escalation Pattern Analysis
• Kaiser R. — Economic Abuse in Coercive Control Trajectories: Applying Escalation Pattern Analysis to Intimate Partner Violence
• Practice Direction 12J — Child Arrangements & Contact Orders: Domestic Abuse and Harm
• Home Office — Controlling or Coercive Behaviour Statutory Guidance
• Re C (Child Arrangements Orders; Domestic Abuse; Child Focused Court) [2026] EWCA Civ 1216
• Re H-N and Others (Children) (Domestic Abuse: Finding of Fact Hearings) [2021] EWCA Civ 448
• Family Procedure Rules Part 25 — Experts and Assessors