KEY TAKEAWAYS
- Violence risk assessment estimates probability under defined conditions. It does not predict an individual's behavior with certainty.
- Structured methods improve consistency, but no score should decide detention, sentencing, discharge or release by itself.
- The most useful assessment explains risk scenarios, changeable factors, protective factors and concrete management options.
A court, hospital or review board may ask whether a person is likely to become violent. The question sounds simple, but it hides several decisions. What behavior counts, over what period, in which setting and under which future conditions?
Updated September 18, 2026: clarified how the HCR-20, actuarial tools and structured professional judgment differ, and strengthened the distinction from behavioral threat assessment.
Violence risk assessment turns that broad concern into a structured clinical and forensic task. Its purpose is not to find certainty. Its purpose is to make the reasoning transparent enough to examine, update and use safely. Start with the forensic assessment framework to compare this referral with competency and criminal-responsibility evaluations.
When the referral concerns an evolving targeted-violence situation rather than a formal individual risk opinion, the relevant model may be a multidisciplinary behavioral threat assessment and management process. The two practices can inform one another, but their teams, authority and operational purpose are not identical.
What violence risk assessment means
In forensic psychology, a forensic risk assessment examines a defined harmful outcome in a legal or justice-related context. The referral must specify the risk being evaluated: violence, for example, is not the same outcome as any reoffending. This article focuses on violence risk; findings from one outcome should not be transferred automatically to another.
Violence risk assessment is the structured evaluation of a person's likelihood of engaging in a defined form of violence within a stated time and context. It combines records, interviews, relevant research and a method suited to the referral question.
The result should be conditional. Risk may look different in a secure hospital, during supervised community treatment or after a return to substance use and unstable housing. A label without those conditions says little about what may happen next.
Assessment is also different from diagnosis. A mental health diagnosis may be relevant, but diagnosis alone does not establish violence risk. Most people with mental illness are not violent, and many pathways to violence do not depend on a psychiatric disorder.
“A useful risk opinion does more than rank concern. It shows what could change the outcome.”
Forensic Psychology editorial principle
Prediction, formulation and management
Prediction asks how likely an outcome may be. Formulation explains how risk could develop for this person in a specific situation. Management identifies steps that may interrupt that pathway.
A strong assessment uses all three. A number without a case explanation is hard to act on, while a story without structured evidence is hard to test.
Violence risk assessment versus behavioral threat assessment
Both practices connect evidence with prevention, but they do not always answer the same referral question. The distinction concerns purpose and setting, not which approach can predict a person's future with certainty.
| Question | Forensic violence risk assessment | Behavioral threat assessment and management |
|---|---|---|
| Main purpose | Develop an individual risk opinion and management recommendations for a specified clinical or legal decision. | Identify and address concerning behavior in an evolving targeted-violence situation. |
| Typical output | A documented formulation with scenarios, limitations and conditions for review. | A coordinated intervention and monitoring plan that changes with new information. |
| Shared limit | A tool score cannot determine what one person will do. | A profile, diagnosis or single concerning act cannot establish a certain violent outcome. |
The Secret Service's threat assessment guidance emphasizes corroborated information and proactive management. Our threat assessment guide explains that process in more detail. An organization may need both approaches, with clear professional roles and lawful information sharing.
Questions a risk assessment should answer
The referral question controls the entire evaluation. “Is this person dangerous?” is too broad to support a careful opinion.
- What type of violence is being considered?
- Who or what might be at risk?
- What time period does the opinion cover?
- In which setting or future scenario?
- Which factors could increase or reduce risk?
- What supervision, treatment or environmental changes may help?
- What information is missing, disputed or uncertain?
These questions make the opinion more specific and easier to revise when circumstances change.
Three approaches to violence risk assessment
Methods differ in how they select information and combine it into a conclusion. The three broad approaches are not interchangeable.
| Approach | How it works | Main strength and limit |
|---|---|---|
| Unstructured clinical judgment | The evaluator selects and weighs information without a formal guide or fixed rule. | Flexible, but reasoning may be inconsistent, difficult to audit and vulnerable to bias. |
| Actuarial assessment | Selected factors are scored and combined through fixed rules derived from a reference sample. | Consistent scoring, but the estimate may not transfer well to a different person, population or decision. |
| Structured professional judgment | An evidence-based guide organizes relevant factors, then the evaluator develops case-specific scenarios, opinions and management plans. | Connects structure with formulation, but quality still depends on training, data and disciplined judgment. |
Structured professional judgment is often shortened to SPJ. It does not mean that a clinician can ignore the guide. The structure shapes information gathering, while the final opinion remains tied to the individual case.
How a forensic violence risk assessment works
No single sequence fits every legal or clinical setting. A defensible evaluation usually includes these connected steps.
- 01
Clarify the referral
The evaluator defines the outcome, time frame, setting and legal or clinical decision. The limits of confidentiality and the evaluator's role are explained.
- 02
Choose an appropriate method
The tool or framework must fit the person's age, population, outcome and context. Familiarity or popularity alone is not enough.
- 03
Gather multiple sources
Interviews are considered beside health records, legal records, incident reports, collateral accounts and other reliable information. Conflicts between sources are documented.
- 04
Identify relevant factors
The evaluator examines historical, current, future and protective factors. The aim is not to collect every negative fact, but to identify information linked to the defined outcome.
- 05
Build risk scenarios
The report explains how violence might occur, toward whom, under what conditions and with what likely warning signs. More than one scenario may be needed.
- 06
Form an opinion and plan
The conclusion states level or nature of concern, uncertainty and practical management options. Observation, instrument result and professional inference remain separate.
- 07
Review when conditions change
Dynamic risk requires reassessment. A major change in symptoms, substance use, support, supervision or access can alter the opinion.
Common tools and the questions they serve
A tool name does not tell you whether it fits the case. Each instrument was designed for a population, outcome and time horizon.
| Tool or framework | Common context | What to remember |
|---|---|---|
| HCR-20 | Adult violence risk in forensic and mental health settings | An SPJ guide covering historical, current clinical and future risk-management factors. |
| VRAG | Violent recidivism in selected adult offender populations | An actuarial approach that depends on the reference group and scoring rules. |
| START | Short-term risks and strengths in mental health settings | Includes dynamic vulnerabilities and strengths across several adverse outcomes. |
| BVC or DASA-IV | Imminent aggression in inpatient settings | Brief tools for short-term monitoring, not long-term community violence prediction. |
| SAVRY | Violence risk in adolescents | An SPJ guide designed for youth and developmental context. |
| SARA | Intimate partner violence risk | Focuses on a specific form of violence and its management needs. |
NICE recommends considering brief structured instruments such as the BVC or DASA-IV rather than unstructured judgment alone for monitoring violence and aggression in inpatient psychiatric settings. That recommendation does not turn a short-term ward tool into a general forensic prediction method. Youth assessment requires separate developmental evidence, as our guide to juvenile forensic psychology explains.
Sexual recidivism is also a distinct outcome with its own instruments, reference groups and interpretation problems. The sexual offense risk assessment guide explains why a general violence tool or one score cannot be transferred automatically to that question.
Static, dynamic and protective factors
Risk factors do not operate like switches. Their meaning depends on timing, combination and context.
Historical or relatively static factors
Past violence, age at first violence and a long pattern of rule violations may inform baseline concern. These facts usually cannot be changed, but they can help identify recurring pathways.
Dynamic factors
Current symptoms, substance use, anger, treatment engagement, relationship conflict, housing and access to potential victims may change. Dynamic factors matter because they can become targets for monitoring and intervention.
Protective factors
Stable support, engagement with care, coping skills, structured daily activity and realistic future plans may reduce risk in a specific setting. Protective does not mean permanently safe. It means the factor may interrupt or weaken a plausible risk pathway.
A factor belongs in the opinion because evidence and the case make it relevant. It should not be included simply because it sounds negative, unusual or socially disapproved.
How accurate is violence risk assessment?
Structured violence risk tools can separate higher-risk from lower-risk groups better than chance, but they cannot determine what one person will do. Accuracy changes with the tool, setting, population, outcome, time frame and threshold.
A 2012 BMJ meta-analysis examined 73 samples involving 24,847 people from 13 countries. For tools predicting violent outcomes, the median area under the curve was 0.72, which indicates moderate discrimination across the included studies.
The same review found a median positive predictive value of 41 percent and a negative predictive value of 91 percent. In plain language, many people placed in the elevated-risk group did not later commit violence, while the tools were more accurate at identifying people who did not offend.
The authors concluded that these instruments may inform treatment and management. The evidence did not support using them as the sole basis for detention, sentencing, discharge or release.
EVIDENCE IN CONTEXT
Why base rates change the answer
Positive predictive value depends partly on how often the outcome occurs in the population being assessed. When serious violence is uncommon, even a useful tool can produce many false positives.
This is why an accuracy figure cannot be copied from one study into every case. The evaluator must consider whether the research sample, outcome and setting resemble the present referral.
From a risk estimate to a useful formulation
A final label such as low, moderate or high can hide the reasoning a decision-maker needs. A formulation shows the likely pathway from trigger to behavior and the places where intervention may work.
It may describe a scenario in which substance use, perceived rejection and loss of housing combine with access to a former partner. The management plan can then address treatment, contact boundaries, housing, monitoring and early warning signs.
Management is not the same as eliminating all risk. It means reducing the likelihood or impact of a defined outcome while checking whether the plan works in practice.
What a management plan may include
- Treatment for symptoms or substance use linked to the risk scenario
- Changes in supervision, contact or access
- Stable housing and structured daily activity
- Monitoring of specific early warning signs
- A response plan for deterioration or loss of support
- Clear responsibility for review and information sharing
Make the review actionable: identify who will follow up, which changes should trigger reassessment and how new information reaches the responsible team. NICE guidance recommends regular review of both the assessment and management plan in its mental health settings; the interval depends on the assessed risk. A report should not imply that yesterday's conditions remain unchanged.
Limits, bias and fairness
Risk assessment can affect liberty, treatment and community access. That power makes transparent limits essential.
Incomplete or unequal data
Records may reflect differences in policing, access to care, institutional response and documentation. More recorded incidents do not always mean more underlying behavior. The evaluator should examine how the data were produced.
Contextual and confirmation bias
A dramatic allegation or an early high-risk label can shape how later information is read. Separating observations from inferences, considering alternatives and documenting disconfirming evidence can reduce this influence.
Population fit
A tool validated mainly in one population may not perform the same way across age, gender, culture, disability or setting. Translation and cultural adaptation involve more than changing the language of an item.
False positives and false negatives
A false positive may restrict someone who would not have become violent. A false negative may leave a preventable risk unaddressed. The acceptable balance depends on the decision, consequences and available safeguards.
Objectivity does not mean pretending values are absent. It means stating the question, evidence, method, uncertainty and consequences clearly enough for others to examine. Our guide to ethics in forensic psychology explains the wider duties behind that process.
A realistic risk formulation example
Consider a person with one serious assault during intoxication five years ago. They have remained stable in a structured setting, completed substance treatment and developed reliable family support. A simple history-based label may keep attention fixed on the old event.
A formulation asks what made that assault possible. The relevant pathway may involve heavy alcohol use, escalating jealousy, access to a partner and failure to seek help when angry. Current treatment and support weaken that pathway, but a return to drinking and unwanted contact could strengthen it again.
The opinion is therefore conditional. Risk may be lower under treatment, sobriety, stable housing and contact boundaries, then rise if several protections fail together. The plan monitors those changes instead of treating the person as permanently safe or permanently dangerous.
How to read a violence risk assessment report
A clear report lets the reader trace the conclusion back to the evidence. Use these questions to test its quality.
- Is the violent outcome defined?
- Are the time frame and setting clear?
- Does the method fit the population and purpose?
- Are records, interviews and collateral sources distinguished?
- Are static, dynamic and protective factors explained?
- Does the report describe plausible risk scenarios?
- Are uncertainty and missing information stated?
- Do management recommendations connect to the formulation?
- Is the final decision left to the proper legal or clinical authority?
A long report is not automatically a strong report. The reasoning should remain specific, testable and understandable.
Common questions about violence risk assessment
What is a violence risk assessment?
A violence risk assessment is a structured evaluation of how likely a person may be to engage in a defined form of violence under specified conditions and within a stated period. It also examines factors that could increase, reduce or help manage that risk. It is a probability-based professional opinion, not a prediction with certainty.
Can a psychologist predict whether someone will become violent?
No psychologist can know with certainty whether an individual will commit violence. A qualified evaluator can estimate risk using case information, research, structured methods and professional judgment. The opinion should state its time frame, scenario, uncertainty and practical management implications.
What is the HCR-20?
The HCR-20 is a structured professional judgment guide used to assess and manage risk for future violence. It organizes historical, current clinical and future risk-management information. It supports a reasoned formulation and management plan rather than producing a stand-alone mathematical probability.
What is the difference between actuarial and structured professional judgment tools?
Actuarial tools combine selected factors through fixed scoring rules to estimate an outcome in relation to a reference group. Structured professional judgment tools guide the evaluator through evidence-based factors and require a case-specific formulation. Both approaches depend on the population, purpose, data quality and correct use of the instrument.
What factors are considered in violence risk assessment?
Relevant factors may include past violence, current symptoms, substance use, treatment response, access to potential victims or weapons, recent losses, stress, supervision, housing and social support. The exact factors depend on the referral question and tool. A responsible assessment also considers protective factors and plausible future situations.
How accurate are violence risk assessment tools?
Accuracy varies across tools, settings, populations, time periods and outcomes. A major meta-analysis found moderate overall discrimination for violence tools, but many people classified at elevated risk did not later offend. A score cannot be treated as certainty or used alone to decide detention, sentencing, discharge or release.
How long is a violence risk assessment valid?
There is no universal expiry date. An assessment describes risk under the conditions known at the time. Major changes in symptoms, substance use, relationships, housing, treatment, supervision or access to victims can require review, especially when the opinion concerns short-term risk.
Who can conduct a forensic violence risk assessment?
The evaluator needs professional competence in the relevant population, legal context, assessment method and local standards. Depending on the setting, the work may be performed by a qualified forensic psychologist, psychiatrist or another trained professional. A degree or license alone does not establish competence with a specific tool or referral question.
Sources
Professional guidance and peer-reviewed reviews were selected for their relevance to forensic practice, structured methods, predictive performance, risk management and bias. Sources were reviewed on September 18, 2026.
- U.S. Secret Service, National Threat Assessment Center. Behavioral Threat Assessment Units (2024).
Official guidance on behavior-based targeted-violence prevention and ongoing management, rather than profiling or certain prediction.
- American Academy of Psychiatry and the Law. Practice Guideline for the Forensic Assessment (2015).
Professional guidance on objectivity, collateral information, legal context and the conduct of forensic risk assessment.
- National Institute for Health and Care Excellence. Violence and Aggression, NG10.
Official recommendations on structured short-term assessment, regular review and risk-management planning in mental health settings.
- Fazel, S. et al. (2012). Use of Risk Assessment Instruments to Predict Violence and Antisocial Behaviour.
Systematic review and meta-analysis of 73 samples involving 24,847 people across 13 countries.
- Ramesh, T. et al. (2023). Violence Risk Assessment Instruments in Forensic Psychiatric Populations.
Systematic review and meta-analysis examining predictive performance in specialist forensic psychiatric populations.
- Douglas, K. S. & Shaffer, C. S. (2021). The Empirical Evidence Base for the Use of the HCR-20.
Review of study designs and the limits of transferring predictive findings into structured professional judgment practice.
- Singh, J. P. et al. (2011). Structured Assessment of Violence Risk in Schizophrenia and Other Psychiatric Disorders.
Systematic review of the validity, reliability and content of structured violence risk instruments.
- Dror, I. E. et al. (2024). A Practical Approach to Mitigating Cognitive Bias Effects in Forensic Casework.
Open-access guidance on contextual bias and practical safeguards for forensic decision-making.