RelationWatch

RelationWatch Research Brief #04

RW-RB-004

When Relationship-Connected Homicide Is Followed by Suicide

Recorded circumstances, relationship context, warning indicators, and available event-sequence evidence in relationship-connected incidents classified as murder-suicide. The victim group can include a partner, children, household members, or other people harmed in the same incident.

Based on a fixed 1284 verified RelationWatch casesPatterns, not causal explanations
136Classified murder-suicideCore analytical group in this fixed snapshot
132Known relationship statusOf the murder-suicide group
25Explicit order in event textHomicide recorded before suicide in available event text
194Broader self-harm signalKept separate because classifications differ

What happened to the other cases?

This Brief analyzes 136 of 1284 verified cases because they are explicitly classified Murder-Suicide. The remaining 1147 cases are not in this core classification. They may include homicide, attempted homicide, injury, disappearance, or other relationship-connected incidents; they are not evidence that suicide, self-harm, or relationship factors were absent. The broader self-harm signal is displayed separately because the current historical fields do not consistently identify who died by suicide, the exact attempt outcome, the exact order, or every child/third-party victim.

Relationship situation → recorded warning → outcome

The Sankey is limited to the strongest structured co-occurrences. Hovering any node or link keeps its connected path visible from relationship situation through the recorded outcome. It is not a complete timeline and does not show that one condition caused another.

Recorded warning indicators

Ranked structured indicators within the 136 cases in this analysis.

Relationship context

Status recorded in the core group; categories describe recorded context, not victim identity.

Available sequence evidence

Existing timeline text is incomplete; this chart separates explicit ordering from ambiguity and missing signal.

Countries represented

Dataset composition only, not a national prevalence estimate.

Recorded motives

Structured motives can overlap and are not proven explanations for suicide.

Age at incident

Recorded-person age distribution among victims and offenders / suspects.

Recorded pathway snapshots

Each row is a documented intersection, not a claimed sequence or causal pathway.

74 cases — Ongoing relationship → Escalation Pattern / Violence → murder-suicide classification
51 cases — Ongoing relationship → Firearm Access / Concerning Weapon Behavior → murder-suicide classification
43 cases — Ongoing relationship → Divorce Pending / Recent Separation → murder-suicide classification
25 cases — Separated / former → Divorce Pending / Recent Separation → murder-suicide classification
23 cases — Separated / former → Escalation Pattern / Violence → murder-suicide classification
21 cases — Ongoing relationship → Threats → murder-suicide classification

How to read this Brief

  • “When” is deliberate. The data support descriptive circumstances and repeated co-occurrences, not a general explanation of why a person died by suicide.
  • Historical records do not consistently identify the person who died by suicide, a nonfatal attempt’s outcome, every additional child or third-party victim, or exact event order.
  • The new structured incident-sequence fields will improve these variables for future cases; they do not retroactively fill missing information in this fixed publication.
  • RelationWatch uses publicly available OSINT sources. Findings describe this fixed dataset, not population prevalence.
All Research Briefs