A Law Enforcement Field Guide to Coercive Control by Carisa Carlton
The 911 call comes in as a “verbal dispute.” The responding officers find a tidy home, a calm man, and a woman who says everything is fine. No visible injuries. No witnesses willing to talk. No crime, apparently. Officers clear the scene in twenty minutes and move on to the next call. Eleven months later, they return to the same address. This time for a homicide.
The scenario above is a composite drawn from patterns documented across domestic violence fatality reviews nationwide, and nearly every veteran officer and probation professional has lived some version of it. But the file on that address tells a different story than any single call ever could: the wife quit her job abruptly, she fell down the stairs, she has no access to their bank accounts, the sister stopped visiting, the husband does all the talking at every contact. The hole in the bathroom door. He just had a bad day, again.
None of it is a crime under the incident-based model. Yet together, all evidence points to the most dangerous behavioral pattern in domestic life: coercive control.
Law enforcement nationwide is trained to respond to physical violence. But the behavioral framework underlying the majority of domestic homicides is not a single violent act, rather a sustained pattern of coercive control. Until officers and probation professionals are trained to see the pattern, we will keep clearing scenes that are, in fact, active crime scenes in slow motion.
Why the Incident Model Fails
American domestic violence response was engineered around the discrete incident: a call, an injury, an arrest decision, a report. That architecture made sense when the reform goal of the 1980s and 1990s was getting police to treat assault in the home as seriously as assault on the street. But this model has a blind spot the size of the problem itself.
Forensic social worker Evan Stark, who co-founded one of America’s first battered women’s shelters, spent three decades documenting that blind spot. In his landmark 2007 book Coercive Control: How Men Entrap Women in Personal Life, Stark showed that in millions of abusive relationships, the primary harm is not physical injury but a pattern of control that manifests as entrapment and more closely resembles hostage-taking than assault.
Stark argued that coercive control should be understood as a “liberty crime”: the theft of a person’s autonomy, one rule and one threat at a time. But even Stark may not have fully articulated the vast variety of power and control strategies. My own work, which builds on Stark’s findings, highlights additional strategies for exerting and maintaining power and control: weaponization of love and affection, spiritual rape of women through lies, religious hierarchy, or even the resulting equalization of power in victimology.
The operational consequence for law enforcement is stark, and it is measurable. Research from the Australian Homicide Project found that fully half of men convicted of killing an intimate partner reported no physical or sexual assault against her in the year before the killing. A University of Gloucestershire analysis of 358 UK domestic homicides led by criminologist Jane Monckton Smith found that controlling behavior was present in 92 percent of the killings. Read those two findings together and the message for patrol and probation is unmistakable: the absence of prior physical violence does not mean the absence of lethal danger. In coercive control cases, the pattern is the warning. Officers and families can analyze a single message for coercive control with a free tool.
If Pattern Is the Warning, Then Learning to Read It Is the Work
Coercive control is a spectrum of overlapping, non-linear tactics aimed at establishing hierarchy and maintaining power and control. Coercive control is not a fixed sequence of behaviors. Any number of tactics can be customized and utilized to produce the desired end result: complete control.
Most of the individual tactics that build a pattern are neither violent, as we traditionally define violence, nor are they illegal in many jurisdictions. Coercive control can be a gradient, but not necessarily a gradient. Consider this real life Newport Beach case, names changed for the purposes of this article:
Steven and Arlene met for a first date at a restaurant for dinner. Afterward, and unbeknownst to her, he followed her home. To this day, it remains a mystery how he skirted the security guard leading into her gated community. Arlene was alarmed to find Steven standing outside her vehicle when she got out of her car. He was just making sure she got home safely, he said.
Later, Steven dropped small gifts at her doorstep, again somehow evading security guards. She found these acts simultaneously flattering and alarming.
Steven often left small personal belongings, such as his sunglasses, at her house, on accident, he said, so he would have to return late at night to fetch them. Later we learned that the real purpose was to make sure Arlene was home alone.
Within two weeks of meeting Arlene, Steven told her she was the love of his life. He wanted her to “be his woman, and to be her man.” He proposed within three months and they had an impromptu fairytale wedding.
Steven did not want her to work, he would provide for the household. Since he was paying, he would decide where they lived and what day the housekeeper came. He did not want her name on the bank accounts, he would control the money, he said. When she was not at home when he arrived from work, he raged. She is a lousy cook, he joked with her family. She did not know a mosquito bite from a gnat bite on her arm, he argued. The woman he is flirting heavily with online is just a friend, stop acting so crazy, stop violating his boundaries by reading his public-facing social media interactions, he told her. During an argument, he forced her out of the car and left her to walk home. He is not ready for a baby, have an abortion, he commanded.
None of these behaviors register as crimes. But friends and family are concerned, and escalating patterns are pointing toward danger.
Six months into the marriage, he pushes her off the bed. He grabs her arm. He punches a hole in the bathroom door she locked herself behind. Finally, he kicks her. She calls the police. But there are no injuries, no bruises, no red marks. His word against hers.
To punish Arlene for calling the police, to regain lost power and control, he robs her. Steven takes all of the money she had saved in the top drawer beside the bed.
Juxtaposed to all this, he buys her expensive gifts. New earrings from Tiffany’s delivered to the front door while he is at work. Flowers to make up for the night before. A weekend vacation to a yoga retreat in the mountains. He loves her, he says. He’s just under stress.
All along, she fights back. She argues. She’s no shrinking violet. But with every psychological blow, she is weakened. Her self esteem is diminished. She is dependent. Financially and emotionally.
There’s been an emergency room visit. But Arlene is afraid to tell the truth about the stairs: she doesn’t want him arrested. He is her husband. She just wants the situation under control.
Outsiders think: Steven and Arlene are the perfect couple. They are gorgeous, rich, powerful, the intellectual elite. Abuse couldn’t possibly be happening in that penthouse. But it happens in every socioeconomic class, every education level, every ethnicity, every culture.
A year later, family ask the police to do a welfare check. She hasn’t responded to texts or calls in four days, they said. Police find her strangled to death.
Each time police came to that address, they were looking at a single incident. But each incident was a code in a larger record.
Anthropologists are trained to identify nuances by assigning qualitative codes across large sets of data, and the same method can work in domestic calls. Following her home from a first date is coded as surveillance and stalking. Slipping past security to leave gifts at her door is coded as disregard for both implicit and explicit boundaries. Refusing to put her name on the bank accounts is coded as financial control. Berating her for questioning his online affairs is coded as gaslighting, and the act itself is triangulation. No single code is an arrest. Read together, across every prior contact with the same person or persons, they are a portrait of escalating danger.
Every generic writing on coercive control suggests the victim is passive, the incidents are extreme, and the pattern is as clear as black and white. But Arlene did fight back. She did argue her rights. She and her family did question what was happening. And yet, the pattern was not obvious as a sign of “danger” to her, outsiders, or police officers.
Coercive control is a pattern. But it’s not a clean, linear, easily identifiable pattern. It can be as confusing to the victim as it is to the police, especially when intermittent affection is present. Most stand-alone coercive control incidents are barely documentable, let alone arrestable offenses.
Not every domestic dispute ends in a death certificate, but police must ask, “What is the pattern?”
The Pattern Is the Predictor
The most important finding in the study of domestic homicide is that coercive control predicts death.
The University of Gloucestershire analysis of 358 domestic violence killings cited earlier found controlling behavior present in 92 percent of them, and half of the intimate partner killers in the Australian Homicide Project had committed no physical assault in the year before they had killed their partner.
Comparing 310 women who were killed against 324 women who were abused but survived, Dr. Jacquelyn Campbell and her colleagues at the Johns Hopkins University School of Nursing identified the factors that most reliably separate abuse from impending homicide. Two of the strongest predictors are not physical acts at all. A partner’s attempt to control the victim’s daily activities is a statistically significant predictor of homicide, and so is a recent or pending separation, which is dangerous precisely because it is the moment the controller loses control.
Stalking follows the same logic. Campbell’s team found stalking, surveillance, and harassment in 70 to 90 percent of the femicides and attempted femicides they studied across a dozen U.S. cities. Separate research places stalking before roughly three out of four intimate partner homicides. Stalking after a breakup is not a nuisance complaint. It is the signature of a controller who has lost access and is escalating.
The pattern of coercive control is not a lesser warning that precedes real danger. In case after case, coercive control is the danger.
This is where the officer’s field reading and the homicide researcher’s data meet. The behaviors that predict death are, to a striking degree, the same quiet, non-criminal tactics of the pattern: surveillance, isolation, control of money and activities, and the refusal to accept that the relationship is over.
The Danger Assessment for Law Enforcement, developed by Campbell and Dr. Jill Messing with the Jeanne Geiger Crisis Center, is a short instrument designed to be used at the scene, and it anchors the Domestic Violence High Risk Team model now operating in jurisdictions across the country. It is worth an honest caveat for the patrol officer: instruments like this were built to forecast physical homicide, and their heaviest weighted items skew toward physical and criminal acts such as prior strangulation, threats to kill, and access to a firearm. Those items matter, but an officer who waits for them is reading late. The controlling pattern is the earlier signal.
There is a second kind of lethality that no homicide instrument was designed to catch, and it is the one coercive control produces most reliably: Coercive control has a significant impact on victims’ minds. A 2023 systematic review and meta-analysis of 45 studies found coercive control moderately associated with post-traumatic stress and depression, on a scale comparable to broader psychological abuse. Gaslighting, degradation, and the steady undermining of a person’s judgment erode self-trust until, as criminologist Sandra Walklate observed, the harm often becomes visible to the victim only after their independence and sense of value are already gone.
Survivors describe becoming a shadow of themselves, unable to trust their own perceptions, frustrated that no one understands, misdiagnosed by medical staff, dismissed by authorities, diminished by systems designed to protect them, and living in constant anxiety. Erosion of self is not a side effect, rather a pathway to self-inflicted death. Research has repeatedly found coercive control more strongly associated with suicidal thinking than other forms of intimate partner violence, with depression as the link between the two, and in the United Kingdom, the National Domestic Homicide Project recently reported a sharp single-year rise in suicides among domestic abuse victims, most of whom were already known to authorities.
Lethality in these cases therefore runs in two directions at once. The controlled partner may be killed by the person controlling them, and the controlled partner may be driven to take their own life. Both outcomes are foreseeable within the same pattern, and neither depends on a bruise. An officer who registers a partner’s visible distress, disorientation, difficulty speaking, and self-doubt at the scene is not observing a fragile personality. They may be witnessing the accumulating evidence of a person in danger from two directions, and recognition is part of the job.
Behavioral Red Flags in Police Reports Before a Domestic Homicide
These are patterns, not single incidents. Each is individually easy to dismiss and rarely arrestable on its own. Read across all prior contacts with the same parties, they form a portrait of escalating danger. A recent or pending separation is not on this list because it is not a behavior; it is a circumstance that raises the risk of every item listed here.
Threats of any kind, including threats to take the children.
Explicit or implicit. Threats to kill, to harm, to report to authorities, to expose, or to take the children. Threatening conduct is a consistent factor across domestic homicide reviews, and threats involving children are a common instrument of control that officers should treat as serious.
Destruction of property.
Punching walls and doors, smashing phones, destroying a partner’s belongings, harming pets. Property destruction is a recognized intimidation tactic, and one large study found that damage to property and pets was more strongly associated with lasting mental harm than physical violence itself.
Minimizing, demeaning, and gaslighting.
Belittling a partner in front of others, denying events the victim witnessed or experienced, and steadily undermining their judgment. These tactics erode a person’s confidence in their own perceptions and are central to how control is established and maintained.
Stalking, following, or surveillance, including electronic monitoring.
Following a partner, tracking their location, monitoring phones and social media, or watching them through devices or networks. Stalking precedes roughly three out of four intimate partner homicides and is one of the clearest signals of a controller escalating after losing access.
The Law Catches Up
For most of American policing history, everything described thus far, short of the final assault, was invisible to the law. That is changing rapidly. In roughly five years, about a dozen states have written coercive control into their statutes, a remarkable pace for a concept that was still largely academic a decade ago. Connecticut, through Jennifer’s Law, and Hawaii were among the first. Arizona, as of this writing, is the most recent.
Nearly all coercive control laws in the United States are civil and family law. They shape protective orders, divorce, and custody decisions, directing judges to weigh patterns of control that once carried no legal weight. They do not, for the most part, create a new offense an officer can arrest for. Hawaii stands as the single exception. It became the first and only state to criminalize coercive control, enacting it in 2021 as a petty misdemeanor within a five-year domestic violence pilot program. Even in Hawaii, however, the tool is narrow. Patterns are difficult to establish at a single call, and the statute’s warrantless arrest authority is still written around physical abuse. Criminalization in Hawaii matters more as a statement than as a patrol instrument.
Arizona shows where the trend is heading. The Alec and Lydia Act (HB 2995), passed June 2026, is the newest and furthest reaching of these laws. Named for two young children killed by their father during court-ordered parenting time that he held despite documented warnings to law enforcement and the family court, the law rewrites custody around the recognition that coercive control endangers children.
Coercive control is not something an officer enforces at a domestic call, rather it is a vivid measure of how seriously the system is taking the pattern. A dozen legislatures naming coercive control as serious harm is not merely a set of new rules. It is a society, and its legal institutions, finally recognizing what survivors and researchers have long known: the pattern is the harm.
What the Pattern Asks of Us
Coercive control is a pattern, not an incident. The damage is in accumulation, in nuance, in a hundred small acts. Each easy to dismiss, together a cage. Coercive control predicts lethal danger more reliably than physical violence alone, and it endangers the controlled person from two directions: the hand of the controller, or their own.
Across the United States, law is beginning to recognize coercive control as danger. What remains is recognition at the door, in the moment, by the officer standing in a living room when nothing fits the incident model.
Recognition is not a checklist that can be applied to every situation, and an officer cannot be tasked with documenting patterns the way a researcher would. But when injuries are absent and the person in front of you cannot articulate plainly what is wrong, the silence itself is information. You may not be able to arrest the pattern, but seeing it changes how you weigh the danger, how seriously you take the next call to that address, and what you flag for the officers who come after you.
The incident model asks: what happened here tonight? The coercive control model asks: what has been happening here all along?
Sources
Stark, E. (2007). Coercive Control: How Men Entrap Women in Personal Life. Oxford University Press. https://academic.oup.com/book/55149
Carlton, C. (2017–present). Selected works on coercive control, power and control, and the study of indigenous and micro-cultures. A decade of applied and ethnographic research underlying this article.
Campbell, J.C., Webster, D.W., & Glass, N. (2009). The Danger Assessment: Validation of a lethality risk assessment instrument for intimate partner femicide. Journal of Interpersonal Violence, 24(4), 653–674. https://pmc.ncbi.nlm.nih.gov/articles/PMC7878014/
Campbell, J.C., et al. (2003). Risk factors for femicide in abusive relationships. American Journal of Public Health, 93(7), 1089–1097. https://pmc.ncbi.nlm.nih.gov/articles/PMC1447915/
Campbell, J.C., et al. (2003). Assessing risk factors for intimate partner homicide. NIJ Journal, 250. U.S. Department of Justice, Office of Justice Programs. https://www.ojp.gov/pdffiles1/jr000250e.pdf
Glass, N., Laughon, K., Campbell, J., et al. (2008). Non-fatal strangulation is an important risk factor for homicide of women. Journal of Emergency Medicine, 35(3), 329–335. https://www.sciencedirect.com/science/article/abs/pii/S0736467907004143
McFarlane, J., Campbell, J.C., et al. (1999). Stalking and intimate partner femicide. Homicide Studies, 3(4), 300–316.
Sheed, A., Brandt, C., & McEwan, T.E. (2024). The relationship between stalking, homicide, and coercive control in an Australian population. Homicide Studies. https://journals.sagepub.com/doi/10.1177/10887679241268032
Jeanne Geiger Crisis Center / Geiger Institute. Danger Assessment for Law Enforcement (DA-LE) and the Domestic Violence High Risk Team Model. https://geigerinstitute.org/da-le/
Blake, N., et al. (2025). Coercive control and intimidation: stronger links to adult mental health than physical violence (Australian Child Maltreatment Study analysis). Summarized via the National Elf Service. https://www.nationalelfservice.net/mental-health/trauma/coercive-control-and-intimidation-stronger-links-to-adult-mental-health-than-physical-violence/
Lohmann, S., et al. (2023). The trauma and mental health impacts of coercive control: A systematic review and meta-analysis. Journal of Interpersonal Violence. https://pmc.ncbi.nlm.nih.gov/articles/PMC10666508/
Wolford-Clevenger, C., & Smith, P.N. (2017). Associations of emotional abuse types with suicide ideation among dating couples. Journal of Aggression, Maltreatment & Trauma, 26(9). https://www.tandfonline.com/doi/abs/10.1080/10926771.2017.1335821
UK National Domestic Homicide Project (2024). Rise in suicides among domestic abuse victims. Summarized via DV-ACT. https://www.dvact.org/post/domesticabuseandsuiciderisk
Walklate, S. (2022), on the delayed visibility of coercive control to victims. Discussed in What Is Coercive Control? Understanding Hidden Abuse. https://womens-safety.com/blog/what-is-coercive-control-understanding-recognizing-hidden-abuse/
Domestic Abuse Intervention Programs (Duluth Model). Power and Control Wheel. https://www.theduluthmodel.org/wheels/
Office of the Federal Ombudsperson for Victims of Crime (Canada). Understanding coercive control in the context of intimate partner violence. Citing Monckton Smith et al. (control present in 92% of 358 UK domestic homicides) and Johnson et al. (2019), Australian Homicide Project. https://www.canada.ca/en/office-federal-ombudsperson-victims-crime/publications/research-recherche/ucc-ccc.html
Hawaii State Legislature. Act 238 (2021), amending HRS § 709-906 and § 586-1 to add coercive control as a petty misdemeanor; five-year domestic violence pilot program established by Act 19 (2020). https://data.capitol.hawaii.gov/hrscurrent/Vol14_Ch0701-0853/HRS0709/HRS_0709-0906.htm
Connecticut General Assembly. Public Act 21-78, “Jennifer’s Law” (2021), incorporating coercive control into civil domestic violence protections.
California Legislature. SB 1141 (2020), Domestic Violence: Coercive Control, amending Family Code § 6320. https://leginfo.legislature.ca.gov/faces/billNavClient.xhtml?bill_id=201920200SB1141
Arizona State Legislature. HB 2995 (2026), the Alec and Lydia Act. Senate fact sheet and adopted text. https://www.azleg.gov/legtext/57leg/2R/summary/S.2995FFL.pdf
University of Rochester School of Nursing (2026). Centennial spotlight: Jacquelyn Campbell contributes to prevention of intimate partner violence. https://son.rochester.edu/newsroom/2026/cs-jacquelyn-campbell.html