Not All Homicide Is Sudden
What Rising Domestic Violence Deaths Reveal About the Coercive Control Timeline
Last week, journalist Justin Klawans reported in The Week that domestic violence crimes have risen 2.5% nationally over four years. The more disturbing pattern, though, is local. In Chicago, overall homicides have fallen by nearly half since 2021, while domestic violence related deaths moved in the opposite direction, rising 15% over that same period. Kansas City, according to the same report, has had only one deadlier year for domestic violence than 2025 in the decade since 2014. Massachusetts recorded a 19% increase in domestic violence killings from 2023 to 2024. The National Domestic Violence Hotline's call volume has grown by nearly 60% since 2020, from roughly 637,000 annual calls to about a million.
Buried in that reporting is a detail that will be familiar to anyone who has sat across from a survivor or reviewed a family court file. Klawans quotes Francine Stark, executive director of the Maine Coalition to End Domestic Violence, who told a Portland television station that victims of domestic violence homicide “very, very frequently” told someone in their life, “I'm afraid that my partner is going to kill me.” Klawans's piece goes on to note that many people “experience abuse and violence and never bring in the criminal system,” which can allow domestic violence crimes to go unpunished. That kind of advance warning is not an anomaly. It recurs often enough in the research on intimate partner homicide that mapping exactly where it fits into the lead-up to a killing has become its own area of study, as the two sources below show.
The news article uses that reporting to make a point about underreporting: many people who experience abuse never bring the criminal legal system in at all. The two studies below take a different approach to the same warning sign. Rather than treating a victim's stated fear as anecdote, they treat it as a measurable data point with a documented place in the progression toward a killing.
The dominant cultural narrative around intimate partner homicide still leans on the language of the sudden: a fight that got out of hand, a “crime of passion,” a snap. Forensic criminologist Jane Monckton Smith's 2020 study in Violence Against Women pushes back hard on that narrative. Drawing on 372 cases of intimate partner homicide, built from interviews with bereaved families and public protection professionals, Monckton Smith mapped what she calls the domestic homicide timeline. It is an eight-stage progression that moves from a pre-relationship history of control, through the early idealization and commitment of the relationship, into the tightening of coercive control, a trigger (frequently separation or the threat of it), escalation, a change in the perpetrator's thinking, and finally planning and homicide. The eight stages are as follows:
Pre-relationship: The perpetrator has a prior history of domestic abuse, controlling patterns, or stalking before entering the current relationship.
Early Relationship: The relationship begins with a highly accelerated pace, marked by rapid commitment, early declarations of love, possessiveness, and jealousy.
Relationship: The relationship becomes characterized by coercive control, where the perpetrator uses monitoring, stalking, and sometimes violence or sexual violence to maintain dominance and establish strict routines.
Trigger/s: An event occurs that threatens the perpetrator's control or status, overwhelmingly triggered by the victim attempting to separate or withdraw her commitment.
Escalation: The perpetrator attempts to re-establish control through an increase in the severity or frequency of abuse, stalking, begging, or threats.
Change in Thinking/Decision: A critical shift occurs where the perpetrator feels they have irretrievably lost control (often described as "last chance thinking") and decides that homicide is the only resolution.
Planning: Evidence of premeditation emerges, which may include buying weapons, conducting internet searches on how to kill, isolating the victim, or organizing finances.
Homicide: The killing itself, which can involve extreme violence, may be accompanied by the perpetrator's suicide, or may involve attempts to cover up the crime.
The motivation to abuse and the motivation to kill are, in her framework, one and the same, because both grow out of a need for control over the other person. What looks, from the outside, like a sudden loss of control is, in the cases she examined, closer to the opposite of that. Rather than losing control, the perpetrator is understood as reacting to an actual or threatened loss of it, and the killing is often the final, deliberate stage of a process that coercive control set in motion long before.
For the people reading this, that reframing changes what “risk” looks like on paper. A landmark 2003 NIJ funded case control study across eleven U.S. cities, the research underlying Jacquelyn Campbell's widely used Danger Assessment instrument, found that a prior threat to kill was associated with nearly a 15-fold increase in risk of intimate partner homicide, alongside factors like access to a firearm in the home and recent separation or estrangement. Considered on its own, a statement like the one victims are already making to friends, family, or coworkers turns out to be one of the more predictive data points documented in this body of research, not simply an unverifiable detail that reads as color in a news story.
For judicial officers, this has direct implications for how protective order hearings, custody evaluations, and bail determinations get read. A petitioner who reports that her partner said he would kill her is not offering hyperbole or an unverifiable subjective fear. She is naming a variable with a documented, quantified relationship to lethal outcomes.
For legal advocates, it means coercive control history, including stalking, monitoring, isolation, and escalating jealousy, deserves to be treated as part of the lethality picture rather than as a separate, softer category of harm that gets addressed only after physical violence has already occurred.
For mental health professionals working with survivors, it offers language for what many already sense clinically: the fear a client describes as “getting worse” or “different lately” often tracks a real, nameable stage in a documented progression, not simply an intensifying anxiety symptom.
For medical professionals, a patient's disclosure that her partner threatened to kill her, made during an intake, a wellness visit, or an emergency department encounter, belongs in the chart as a documented risk factor rather than a passing remark and, in many cases, requires mandatory reporting. Asking about that history, and recording it clearly, gives the next clinician who sees her, or law enforcement and legal/mental health professionals, a piece of information with real weight behind it.
None of this predicts that any individual case will end in homicide, however. Monckton Smith's model, like Campbell's risk factors, was built to improve pattern recognition, not to generate a certainty score. But the rise reported by The Week is not actually a mystery. It is what happens when coercive control escalates as an unrecognized, unnamed, and unaddressed issue until it reaches the final stage of a process that researchers have already mapped.
The data from Chicago, Kansas City, and Massachusetts describe outcomes. The Monckton Smith timeline, and the risk factors Campbell's team identified two decades ago, describe the path that gets you there, and the points along it where recognition, documentation, and intervention are still possible. The missing link is the education needed for the professionals most likely and best situated to intervene.
Sources:
News story: The Week: “Domestic violence killings are rising even as homicide rates fall”
Supporting data: Campbell et al. (2003), “Assessing Risk Factors for Intimate Partner Homicide,” NIJ Journal
