Protective Factors in Violence Risk Assessment and Risk Management
Author: Ukrainian Psychological Hub · Published: October 11, 2026 · Editorial Policy
What Are Protective Factors in Violence Risk Assessment?
Protective factors in violence risk assessment are characteristics, resources, capabilities, relationships, and conditions associated with a lower likelihood of specified violent outcomes in a defined context. Their practical purpose is to help forensic professionals understand how safety can be sustained, which resources are available, and which conditions require active management. A protective factor may concern an individual’s behavior, a supportive relationship, effective treatment engagement, a predictable environment, or reliable access to services. These factors are evaluated alongside risk-relevant information rather than in isolation. [Burghart] [CDC]
A person can have serious historical risk markers and meaningful current protections at the same time. A history of violence remains historically true after employment, support, or treatment improves; the new circumstances may nevertheless change how future opportunities, pressures, and response options are organized. The assessment therefore asks what may happen, to whom, under which conditions, within what period, and what observable safeguards could alter that trajectory. This is an evidence-informed formulation for risk management, not a declaration that violence will or will not occur. [APA] [Lancet]
The most defensible reading of the evidence is measured. Research supports the relevance of protective factors, particularly when they are assessed with structured methods. Evidence that scores predict an absence of violence is stronger than evidence that deliberately increasing a particular score necessarily prevents violence. Predictive association, incremental validity, intervention effectiveness, and legal justification are separate questions requiring separate evidence. [Burghart] [OShea]
Why Protective Factors Matter to Forensic Decisions
Risk-oriented assessments often inventory prior offenses, instability, threats, substance-related problems, or failures of supervision. Such information helps identify hazards but may leave a practical gap: the decision-maker still needs to know how a person's environment and functioning could become safer. A protective-factor assessment helps fill that gap by examining whether particular strengths are dependable, accessible, relevant to the anticipated violence scenario, and likely to remain available during stress. [Burghart] [START]
This approach improves the specificity of management planning. Instead of a report stating only that a person has elevated concerns, a structured account may describe a reliable support contact, demonstrated conflict-disengagement skills, predictable medication and treatment access when clinically indicated, or a stable housing arrangement that reduces recurring confrontation. Each claim must be documented and tied to a plausible mechanism. A favorable biography, a pleasant interview, or a high global strengths score cannot substitute for evidence that the relevant protection works in the relevant setting. [APA] [Burghart]
Protective information also has an ethical function. Reporting both vulnerability and capability can improve the accuracy and usefulness of a forensic formulation while avoiding a permanent identity of dangerousness. It gives professionals concrete conditions to preserve and expand, without treating a clinical diagnosis or a past justice-system contact as an individual verdict. [APA]
Protective Factor, Strength, Absence of Risk, and Intervention: Four Different Things
A protective factor is an evidence-relevant condition
A protective factor is operationally meaningful when its presence can be described, assessed reliably, and related to a defined outcome. The broader population association may be established even when its causal role in one person's life remains uncertain. Supportive family relationships, for example, can have beneficial associations in youth-violence research, but a specific family relationship requires assessment of its quality, continuity, capacity, and actual effect on foreseeable conflict. [CDC] [Sijtsema]
A strength is a capability that may become protective
Strength describes something an individual or community can draw upon, such as emotional regulation, occupational skills, social connection, or persistence with treatment. It becomes relevant to a violence formulation when there is a credible path from that capability to safer behavior. A strong work ethic may aid routine and stability; its value for a particular threat scenario depends on how the work environment operates and whether the ability persists under stress. Instruments differ in whether their items measure positively demonstrated resources or merely the absence of problems. [SAVRY] [Burghart]
The absence of a risk factor does not always prove protection
No recorded substance misuse, no recent disciplinary incident, or no active symptom does not automatically demonstrate an independent protective process. Some instruments code the low end of a risk dimension as strength, while other instruments intentionally assess a separate positive capacity. This conceptual distinction matters because absence of an adverse event may reflect limited opportunity, incomplete observation, or a genuinely improved condition. Research on youth tools has questioned whether binary protective items sometimes capture the absence of vulnerabilities better than affirmative strengths. [SAVRY]
An intervention is a planned action with an outcome to test
An intervention may aim to build protective capacity through housing support, skill development, coordinated services, or relational repair. The target, mechanism, and evaluation plan should be stated beforehand. A higher protective-factor score after intervention can indicate change, but the score alone does not establish that the intervention caused an improvement in safety. Natural recovery, environmental restrictions, differing opportunities for violence, and measurement error remain alternative explanations. [Burghart] [Duan]
Types of Protective Factors: Individual, Relational, and Environmental
Individual capacities and behavioral skills
Relevant individual protections may include demonstrated problem-solving, the ability to recognize escalating arousal, nonviolent conflict resolution, self-regulation, insight into personally relevant high-risk situations, and the willingness to use an agreed support plan. Evaluators should favor observed examples over abstract endorsements. For instance, a person who has repeatedly exited confrontations without escalation provides different evidence from someone who simply states that they are calm. A skill may be present in structured treatment but unavailable during intoxication, sleep deprivation, humiliation, or acute interpersonal conflict. [CDC] [START]
Personal agency is context-dependent. Someone may manage frustration successfully at work yet struggle during a particular relationship conflict. Protective formulations identify where a capacity has been demonstrated, where generalization remains uncertain, and whether its operation relies on prompts, supervision, treatment continuity, or a predictable setting. Every relevant factor should be described with sufficient behavioral detail that a second professional can audit the judgment. [APA]
Supportive relationships and social bonds
Social connection can provide belonging, early warning, practical assistance, feedback, and alternatives to violent responses. Prosocial peers, caregivers, mentors, and reliable service providers can help reduce isolation and support safer routines. Longitudinal youth research and public-health guidance identify social support and positive adult relationships as relevant correlates of better outcomes. The finding remains population- and development-specific, so it should be applied carefully in adult correctional and forensic psychiatric settings. [CDC] [Sijtsema] [Kleeven]
The existence of a relative or partner is not enough. Practitioners examine whether the relationship is safe, voluntary, noncoercive, realistic, and available during relevant stressors. A family member who is fearful, exhausted, or unable to provide transportation should not be assigned a protective role simply because the person lives nearby. Risk management must safeguard possible victims and must avoid shifting institutional obligations onto unpaid supporters without their informed participation.
Housing, employment, education, and meaningful activity
Stable housing, structured daily activity, educational participation, and feasible employment may support routine, reduce exposure to recurring conflict, and facilitate contact with services. These are hypotheses to examine in the person's actual circumstances, not universal guarantees. A work role that produces intense conflict or a housing arrangement near a threatened person may increase problems despite appearing positive on a checklist. Access, quality, stability, and fit matter more than a generic label. [WHO] [CDC]
A prospective management plan specifies what makes the environment workable: whether housing is secure, whether transport exists, whether work schedules align with treatment, who can respond to a breakdown, and what contingencies can prevent escalating conflict. Structural resources are part of the formulation. Poverty, discrimination, and unavailable services should not be recoded as inherent psychological dangerousness. [APA] [CDC]
Engagement with services and health care
When clinically appropriate, sustained access to mental health or substance-use care can support symptom management, crisis response, continuity, and coordination. Treatment engagement must be evaluated in relation to actual benefit, accessibility, side effects, preferences, and barriers. Attending appointments alone is a process indicator; it does not prove reduced violence. Likewise, the presence of a psychiatric diagnosis cannot be used as shorthand for future violence. [APA] [Lancet]
A comprehensive plan may include relapse-response contacts, voluntary safety arrangements, timely reassessment after material changes, and protection of the person's rights. Health care serves the person's clinical needs; forensic involvement adds a separate question about risk-relevant functioning. These roles require clear consent, disclosure of confidentiality limits, and careful management of competing duties. [APA]
How Do Protective Factors Work? Three Competing Explanations
Direct or promotive association
Under a direct-effect model, greater availability of a protective factor is associated with lower violence rates across levels of other measured risks. For example, strong supportive relationships may correlate with fewer violent outcomes in longitudinal samples. Such an association does not automatically reveal the mechanism: the relationship may have a genuine effect, reflect earlier developmental advantages, or proxy unmeasured resources. Prospective designs and careful confounder control provide stronger support than a retrospective file review. [Sijtsema] [CDC]
Buffering or interaction
A buffering model proposes that a protective factor changes the impact of a particular risk condition. A conflict-resolution skill may be especially consequential when confrontation is likely; a reliable crisis service may matter most during acute deterioration. Statistically demonstrating buffering requires testing an interaction, not merely observing that protective scores and risk scores have opposite correlations with outcomes. Many available studies have limited power to distinguish interaction from an ordinary main effect. [SAVRY]
Change in exposure, opportunity, or response
A third explanation is ecological: protective conditions may alter opportunities for conflict, access to weapons, the availability of immediate help, or the consequences of early warning signs. In this account, safety may improve because situations become more manageable, even if a relatively stable psychological trait remains unchanged. The same condition can lose protective value when context changes. Environmental restrictions can also lower observed events without measuring internal change, so researchers must report supervision intensity and actual opportunity for violence. [Lancet]
Different mechanisms may operate together. A successful assessment therefore identifies the proposed pathway for each major protection, notes observations supporting it, and specifies what future evidence would challenge the judgment. This makes the formulation revisable rather than turning a strengths inventory into an immutable category. [APA]
How Protective Factors Are Assessed in Practice
Forensic protective-factor assessment generally begins with a clear referral question and specified outcome: physical assault in an inpatient unit, interpersonal violence after release, violent recidivism over a defined interval, or a context-specific threat. These outcomes have different base rates, reporting systems, and implications. The evaluator reviews records, conducts interviews when appropriate, considers collateral evidence with proper safeguards, and distinguishes established facts from uncorroborated claims. [APA] [Lancet]
Information is interpreted through a structured method suited to the population and setting. An assessment documents historical risk, potentially changing concerns, existing strengths, situational constraints, foreseeable scenarios, and management needs. Protective ratings can complement a risk-focused instrument or be integrated into a structured professional judgment approach. The tool's manual, training requirements, and validated use determine what a score can reasonably mean. A clinician's unsupported impression that someone is 'low risk because they have support' is not a validated method. [Burghart] [OShea]
A practical case formulation addresses six linked questions: which harmful outcome is at issue; what would make it more likely; which protections are currently demonstrated; which of those directly address the relevant scenario; what could undermine the protections; and who is responsible for monitoring and responding. These are elements of professional reasoning rather than an independent instrument or mechanical scoring formula.
SAPROF: Structured Assessment of Protective Factors for Violence Risk
The SAPROF family was designed to complement risk-oriented assessments through systematic consideration of protective factors. Adult, youth, and other context-specific versions differ and are not interchangeable merely because they share a name. The adult SAPROF has been examined in forensic psychiatric settings and is commonly interpreted in conjunction with structured violence risk approaches. Its purpose includes identifying strengths relevant to clinical and risk-management planning, not producing an unconditional certificate of safety. [Burghart]
In a 2023 meta-analysis covering 39 articles and 5,434 participants across 16 countries, the SAPROF family showed good interrater reliability and moderate-to-good predictive performance for the absence of recidivism and institutional misconduct. Across the combined literature the instruments demonstrated incremental predictive validity when considered with risk-focused assessment. Changes in SAPROF scores were also associated with lower subsequent violent and general recidivism after adjustment for baseline risk. These are important results, with substantial caveats. [Burghart]
Most included research was retrospective, participants were predominantly male, and included studies did not report calibration statistics. Incremental validity in pooled evidence does not mean that every local application improves prediction. A smaller pseudo-prospective investigation of 75 male secure inpatients, for example, found significant prediction of nonviolence but did not establish incremental validity of SAPROF over an HCR-20V3-based assessment in its regression analysis. Practice should report this heterogeneity rather than present one global accuracy claim. [Burghart] [StudyHCR]
START: Strengths and Vulnerabilities in Short-Term Assessment
The Short-Term Assessment of Risk and Treatability, or START, considers dynamic factors in two related directions: as strengths and as vulnerabilities. It is designed for short-horizon judgments across several potential adverse outcomes, including violence toward others, in settings for which the instrument has been developed and evaluated. The two dimensions are valuable because the same broad life domain may contain a genuine capacity and a meaningful vulnerability, each requiring its own appraisal. [START] [OShea]
A 2014 systematic review and meta-analysis identified 23 studies for qualitative review and nine studies with 543 participants for meta-analysis. It reported promising prediction for several aggressive outcomes, while finding mixed or insufficient performance for other target outcomes, limited evidence about ratings over time, and only three meta-analyzed studies judged at low risk of bias. The review also noted that a START-based intervention in one forensic outpatient study did not outperform standard care for reducing violence. The instrument's inclusion of strengths should not be taken as proof that using it improves treatment outcomes. [OShea]
SAVRY and SAPROF-YV in Youth Assessment
The Structured Assessment of Violence Risk in Youth, or SAVRY, considers multiple developmental risk dimensions and protective factors. The SAPROF Youth Version was developed to investigate protective resources with greater emphasis. Youth evaluation requires attention to age, development, family and school context, and the differences between adolescent and adult behavior. A protective process that matters for younger teenagers may function differently for older youth or adults. [Kleeven] [Sijtsema]
In a Dutch sample of 354 justice-involved adolescents and young adults aged 16–26, retrospective SAVRY and SAPROF-YV ratings showed moderate predictive validity for later offending; SAPROF-YV added information beyond SAVRY, with stronger validity among younger individuals. A separate study of 233 young male offenders found small-to-moderate effects across instruments and did not find incremental validity of SAPROF-YV for violent reoffending over competing tools. These contrasting findings are scientifically informative: predictive value depends on sample, outcome, design, and local application. [Kleeven] [YouthCompare]
A longitudinal community study of 963 adolescents found that a social-support factor derived from SAVRY-related data added information about later violence beyond an identified history/dysregulation factor and sex. Another study involving 156 youth on probation found that SAVRY and YLS/CMI protective scores modestly predicted future charges, with AUC values near 0.61 and 0.60, yet did not add predictive validity over risk factors. Group findings establish a useful research direction while limiting claims that any one protection score can resolve an individual justice decision. [Sijtsema] [SAVRY]
Predictive Validity: What Does the Evidence Actually Show?
Predictive validity asks whether an assessment obtained before an outcome can distinguish later cases with and without that outcome. For protective factors, the outcome may be framed as the absence of institutional violence, absence of rearrest, absence of reconviction, or a reduction in violent incidents. These endpoints are related but not equivalent. Rearrest depends on policing and reporting, reconviction depends on legal process, and inpatient incidents depend on observation and institutional recording practices. [Burghart] [Lancet]
An area under the receiver operating characteristic curve, or AUC, measures rank discrimination. An AUC of 0.70 means that for a randomly selected person who later had the outcome and one who did not, the model ranks the outcome case higher roughly 70 percent of the time, assuming the model and outcome direction are specified correctly. It does not mean that a person with a particular score has a 70 percent probability of violence, nor that 70 percent of individual forecasts are correct. [Lancet]
The broader 2023 systematic review of post-discharge violence-risk instruments included 50 publications, 36 tools, and 10,460 participants. It reported mixed performance, pooled AUC estimates of roughly 0.64–0.72 for selected common measures, and high risk of bias in 49 of 50 studies. This review was not a dedicated protective-factor meta-analysis; it provides essential background on the limits of applying forensic predictive measures in practice. [Lancet]
Incremental Validity: Do Strengths Add Information Beyond Risk Scores?
Incremental validity means that protective-factor information improves a model or structured judgment after accounting for the information already available from a risk-oriented measure. A protective scale can correlate with nonviolence yet add little to predictions if it substantially duplicates other information. Demonstrating increment requires explicit comparison of otherwise similar approaches, preferably on independent samples, with attention to statistical optimism and clinically useful differences. [Burghart] [StudyHCR]
The 2023 SAPROF meta-analysis provides evidence of incremental predictive value at an aggregate level. However, studies in youth and adult inpatient samples have reported weaker or absent gains. It is therefore necessary to ask not only whether a test statistic reaches conventional significance but whether the added information changes decisions beneficially at realistic thresholds. Incremental discrimination, improved probability calibration, more actionable plans, and better actual safety outcomes are distinct endpoints. [Burghart] [YouthCompare] [SAVRY]
Calibration, Base Rates, and False Positives
Calibration concerns whether estimated event probabilities match observed event rates among comparable people over a specified period. A tool can sort people moderately well while assigning inaccurate absolute probabilities. The SAPROF meta-analysis noted that studies did not report calibration indices; the broader forensic psychiatry review similarly found calibration infrequently examined. Users should resist converting a favorable protective score into a numerical statement of personal safety unless local, appropriate, calibrated evidence supports that interpretation. [Burghart] [Lancet]
Consider a hypothetical population of 1,000 people with a future violent-outcome rate of 5 percent during a defined follow-up. Even an illustrative test with 80 percent sensitivity and 80 percent specificity would identify 40 true positives among the 50 outcome cases, yet incorrectly flag 190 among the 950 outcome-free cases. Its positive predictive value would be 40 divided by 230, or about 17 percent. These numbers are a mathematical demonstration of base-rate effects, not validated performance estimates for SAPROF, START, SAVRY, or any particular tool.
The consequences differ by context. A false positive may lead to unnecessary restrictions, stigma, detention, or missed opportunities; a false negative may leave a person or potential victim without adequate protection. Risk management itself can prevent an anticipated outcome, creating ambiguity about whether a seemingly false alarm was an error or whether intervention succeeded. Evaluations need to report what management took place, the decision threshold, the actual observation period, and the consequences of mistakes. [Lancet] [APA]
Population Limits, Cultural Fairness, and Transportability
An instrument evaluated mainly among men in secure forensic hospitals should not be assumed equally accurate for women, adolescents, community samples, people with different cultural backgrounds, or populations in other justice systems. Validation requires attention to referral patterns, prevalence, translated items, setting, legal practices, and outcomes. The major SAPROF synthesis identified predominantly male, retrospective samples; broader forensic risk literature also documents major gaps in female-only and lower-income-country evidence. [Burghart] [Lancet]
Socioeconomic conditions can be confused with personal deficits when opportunity and access determine how easily someone can demonstrate conventional strengths. Housing insecurity may reflect an unavailable rental market, rather than impaired motivation. Lack of formal employment may reflect disability, caregiving, discrimination, or immigration restrictions. Assessment must document these conditions and avoid mechanically penalizing people for barriers outside their control. [APA]
For multilingual, neurodivergent, disabled, or culturally diverse examinees, evaluators should investigate whether communication style, support networks, institutional expectations, and access to services change the meaning or measurement of an item. Fairness evaluation ideally addresses subgroup discrimination, calibration, error rates, and the consequences of downstream decisions, not only whether each group receives the same questionnaire. [APA] [Lancet]
Does Increasing Protective Factors Actually Prevent Violence?
There is a scientifically important gap between predicting an outcome and producing it. Longitudinal changes in a strengths score may forecast later nonviolence, but a causal claim requires evidence that deliberately modifying the factor changes violence outcomes compared with a suitable alternative explanation. Randomized studies, strong quasi-experiments, and carefully designed longitudinal studies can contribute. Evaluation should examine actual harmful outcomes and well-being alongside scores, with enough follow-up to distinguish temporary compliance from durable change. [Burghart] [Duan]
The evidence supports reasonable attention to strengths in treatment planning, but it has not established that maximizing a checklist score is a universally effective violence-prevention program. START research illustrates this boundary: promising assessment performance did not entail superiority over standard care in one tested intervention. Correctional intervention studies also vary in quality, implementation, population, and treatment setting. Clinicians should state the status of their proposed mechanism and evaluate it over time. [OShea] [Duan]
From Assessment to a Working Risk Management Plan
Step 1: Specify the violence scenario
Define the outcome concretely. 'Violence' may mean a physical assault in a unit, threats escalating to targeted harm, assault against a known person after release, or a new violent conviction. Describe the relevant time horizon, setting, likely triggers, and who could be affected. Without this specification, a list of strengths has little practical value. Scenario work must rest on actual case information and professional standards, not cinematic profiles or speculative certainty. [APA]
Step 2: Identify proven and fragile protections
For each proposed protective factor, distinguish direct evidence, self-report, collateral report, and professional inference. Record whether the resource is available now, whether it was demonstrated in similar circumstances, how long it has been stable, and what could cause it to disappear. A supportive mentor may be dependable but unavailable overnight; an employment arrangement may end abruptly; effective coping may weaken during intoxication or conflict. Fragility is relevant even when overall functioning is good.
Step 3: Match a protection to the mechanism of concern
A protection becomes more actionable when it addresses the risk pathway at issue. If conflict intensifies when alcohol is used, access to supportive substance-use care and concrete responses to recurrence may be relevant. If the threat occurs during disputes over contact, a professionally coordinated communication and safety plan may matter more than a general measure of sociability. The match should remain a reasoned hypothesis, revised against observed evidence.
Step 4: Share duties and preserve safeguards
Assign named professional responsibilities, available services, monitoring arrangements, and escalation routes. Respect privacy and confidentiality limits, and avoid requiring a potential victim, child, or unsupported family member to manage danger. The person assessed should understand, as far as appropriate, what is expected, how decisions may affect them, and what evidence will be used for review. Plans must be feasible within the jurisdiction and service system. [APA]
Step 5: Set reassessment triggers and outcome measures
Reassessment may be warranted after a change in housing, service access, acute stress, major relationship disruption, release conditions, or behavior indicating escalating concern. Define how the plan's effectiveness will be evaluated: documented conflict episodes, reliable service access, response to warning signs, and relevant violence outcomes. Record both improvements and failures. A safer pathway is sustained by observable practices and feedback, not by assuming that a favorable score remains valid indefinitely. [APA] [Lancet]
Illustrative Case Formulation: How the Same Protective Factor Can Change Meaning
Imagine an adult under community supervision after a violent offense. The historical record indicates previous assaults during repeated interpersonal disputes. At assessment, the person has stable accommodation, regular appointments, and a sibling willing to help with transportation. These details may support planning, but none alone shows that risk has been eliminated. The relevant inquiry is whether dispute-related triggers remain, how the person now responds, and whether support can function during a crisis.
Suppose direct observation and collateral records show several recent instances of leaving an escalating encounter and seeking assistance. That evidence is stronger than a verbal assertion of self-control because it is specific, repeatable, and relevant to the anticipated scenario. If the sibling's availability then changes or the housing agreement breaks down, the earlier protective formulation must be revised. The appropriate conclusion is conditional: observable capacities and resources may contribute to safer behavior while they remain accessible and effective.
This is a hypothetical educational example. It is neither a risk assessment of a real person nor an empirical demonstration that one protective factor caused nonviolence. An actual forensic report would need collateral corroboration, attention to competing explanations, an explicit legal referral question, and a professional judgment grounded in the applicable population and jurisdiction.
Protective Factors and Specific Forensic Settings
Inpatient forensic hospitals
In secure units, observation density, clinical support, rules, group composition, and immediate staff response strongly influence recorded violence. Strengths may include demonstrated de-escalation, effective requests for help, engagement in personally useful activities, and reliable therapeutic relationships. Yet the restricted environment also limits opportunities for certain harmful behaviors; low incident counts there cannot be transferred automatically to an unrestricted community. The observational context is part of the finding. [Lancet] [OShea]
Community supervision and reentry
After release, protections may include suitable accommodation, legal livelihood, coordinated appointments, safe social networks, and timely crisis assistance. The assessment should account for actual access: a referral with a six-month waiting list may offer little immediate protection. Forensic practitioners should avoid interpreting predictable service-system failures as personal moral defects. An effective plan evaluates achievable resources and monitors disruptions, consistent with professional obligations and the specific terms imposed by the court. [APA] [Duan]
Adolescent justice settings
Youth strengths are embedded in changing developmental systems. Caregiving, educational belonging, trusted adults, peer opportunities, and access to structured activities may change markedly over months. Youth risk and protection instruments require age-appropriate evidence and should support services rather than harden transient developmental traits into permanent labels. Even where statistical prediction is moderate, individual legal decisions must be grounded in the governing standards and safeguards. [Kleeven] [Sijtsema] [CDC]
Targeted violence and threat assessment
In threat-related cases, a supportive relationship or treatment contact can matter, but decision-making also depends on credible behavior, escalation, access to means, grievances, situational changes, and safeguards for potentially affected people. A protective-factor total cannot replace a time-sensitive multidisciplinary response to a specific credible threat. The purpose of this article is to explain protective factors; comprehensive threat assessment procedures warrant a distinct evidence-based treatment.
The Legal Significance of Protective Factors
The psychological question is whether a protective condition is supported by evidence and relevant to a defined outcome. A forensic opinion is the professional synthesis offered to a court or agency within an authorized role. A legal standard specifies what a decision-maker must decide under applicable law; a legal finding is the resulting adjudication. These are separate levels of analysis. A score indicating more protection cannot establish criminal innocence, competency, lack of legal responsibility, or eligibility for release by itself. [APA]
U.S. state and federal systems, as well as systems outside the United States, use different statutes, rules of evidence, procedures, and criteria for detention, supervision, civil commitment, and release. The same psychological observation may be relevant to different legal questions in different settings. Reports should therefore specify the referral question, the governing context supplied by counsel or the court, and the scientific limits of the opinion. Courts make legal decisions; experts communicate defensible psychological evidence. [APA]
Ethics, Transparency, and Human Rights
Protective-factor assessment must be competent, impartial, and transparent. The American Psychological Association's forensic specialty guidelines emphasize scientific foundations, appropriate methods, relevant differences among examinees, and clear disclosure of limitations. An assessor should describe sources and missing records, instrument version and purpose, uncertainty, relevant cultural context, alternative interpretations, and the basis for judgment. Promoting rehabilitation is compatible with public safety when claims remain testable and proportionate. [APA]
Assessors should avoid rewarding superficially conventional life arrangements while undervaluing disability accommodations, culturally different support systems, or nontraditional yet effective resources. If a protective claim depends on another person, that person's safety, wishes, capacity, and confidentiality deserve consideration. Any proposed restriction should be connected to an authorized legal process rather than inferred from a numerical score.
Common Myths and Misconceptions
Myth: A large number of protective factors cancels out a serious history. Reality: Historical facts remain relevant. Protection changes the current formulation when it addresses a credible pathway and is supported by observation; there is no universal arithmetic exchange rate between a past offense and a present strength. [APA]
Myth: A SAPROF score predicts exactly who will be nonviolent. Reality: Research evaluates discrimination at group level, and the major meta-analysis identified insufficient calibration reporting. Individual outcomes remain uncertain, and an instrument's performance depends on population and context. [Burghart]
Myth: Any positive personal trait is protective. Reality: A strength must be evaluated for its relevance, stability, practical availability, and actual role in a particular scenario. Generic favorable impressions have limited forensic value.
Myth: Strong predictive validity proves that adding strengths-based services prevents violence. Reality: Prediction research and intervention-effectiveness research answer different questions. Direct evidence of beneficial intervention effects remains more limited than evidence linking many protective scores to favorable outcomes. [OShea] [Burghart]
Myth: People with psychiatric diagnoses should automatically be categorized as violent. Reality: Diagnosis describes clinical criteria, while criminal behavior and violence depend on heterogeneous individual and contextual pathways. Ethical forensic assessment examines the actual referral question without substituting a label for evidence. [APA]
Frequently Asked Questions
What is the simplest definition of a protective factor in violence assessment?
A protective factor is a documented characteristic or circumstance associated with a reduced likelihood of a specified violent outcome, considered in its proper population and context. Its practical significance depends on how it can operate in the individual case. [CDC] [Burghart]
What examples are most common?
Examples considered in professional practice include dependable positive relationships, demonstrated conflict-management skills, stable living arrangements, meaningful daily activities, and appropriate access to services. Their presence on a list is only the beginning; reliability and relevance must be evaluated. [CDC] [START]
Can a person have both high risk and strong protections?
Yes. Historical and current concerns can coexist with substantial resources and demonstrated strengths. A defensible assessment explains the interaction, the limits of each observation, and the conditions under which risk management may work. [Burghart]
What is SAPROF used for?
SAPROF is a structured approach for assessing protective factors that generally complements a risk-focused forensic assessment. It can inform case formulation, discussion of strengths, and management planning; its results are not an unconditional prediction or legal decision. [Burghart]
How is START different from SAPROF?
START examines short-term strengths and vulnerabilities across multiple adverse outcomes in a structured professional judgment framework. SAPROF focuses specifically on protective factors and is commonly paired with a risk-oriented tool. Each has separate manuals, intended contexts, and validation literatures. [OShea] [Burghart]
Do protective factors eliminate the possibility of violence?
No observed protection provides absolute assurance. Relevant resources may reduce risk, change opportunities, or support safer responses, and their effectiveness can vary with circumstances and time.
Can scores determine whether a defendant should be released?
A court or authorized body applies the governing jurisdiction's legal standards. Psychological measures can provide bounded evidence, but their scores do not replace individualized evidence, procedural safeguards, or the legal decision-maker's responsibility. [APA]
What does an AUC of 0.70 mean?
It is a measure of ranking performance across people with and without an observed outcome. It is not a 70 percent probability that a specific person will commit violence and not a guarantee of correct classification for 70 percent of people. [Lancet]
What do false positives and false negatives mean?
A false positive is a high-risk classification followed by no observed target outcome; a false negative is a low-risk classification followed by the target outcome. Interpretation also requires knowing whether active management prevented harm and whether the outcome could be observed. [Lancet]
Are these tools equally valid across cultures and genders?
No such equivalence can be assumed. Evidence varies substantially across gender, age, language, national context, setting, and base rates. Local validation, appropriate interpretation, and attention to structural opportunity and fairness are necessary. [Burghart] [APA] [Lancet]
What should be monitored after a protective-factor assessment?
Professionals should monitor whether protections remain accessible and effective, whether risk-relevant circumstances have changed, whether interventions are implemented as intended, and whether concrete violence-related outcomes or warning signs appear. Reassessment is linked to meaningful changes rather than a fixed assumption of safety.
Conclusion
Protective factors make violence risk assessment more informative when they are anchored in evidence, integrated with a clearly defined violence scenario, and translated into feasible management actions. SAPROF, START, and youth instruments contribute useful structured approaches, with a growing evidence base for predictive association and some incremental validity. Their limitations remain substantive: samples are uneven, intervention effects need stronger tests, local calibration is rarely reported, and decisions have different legal and ethical consequences. The most responsible practice combines rigor about risk with equally rigorous attention to the conditions that can support safety. [Burghart] [OShea] [APA]
For broader context, see Violence Risk Assessment: What It Can Predict and Where It Fails.
Related Articles
References
American Psychological Association (2013). Specialty guidelines for forensic psychology. American Psychologist, 68, 7–19. https://doi.org/10.1037/a0029889
Assessing protective factors for adolescent offending: A conceptually informed examination of the SAVRY and YLS/CMI (2018). Assessment. https://pubmed.ncbi.nlm.nih.gov/29716397/
Burghart, M., de Ruiter, C., Hynes, S. E., Krishnan, N., Levtova, Y., & Uyar, A. (2023). The Structured Assessment of Protective Factors for violence risk (SAPROF): A meta-analysis of its predictive and incremental validity. Psychological Assessment, 35, 56–67. https://doi.org/10.1037/pas0001184
Centers for Disease Control and Prevention (2024). Risk and protective factors: Youth violence prevention. https://www.cdc.gov/youth-violence/risk-factors/index.html
Duan, W., Wang, Z., Yang, C., & Ke, S. (2023). Are risk-need-responsivity principles golden? A meta-analysis of randomized controlled trials of community correction programs. Journal of Experimental Criminology. https://doi.org/10.1007/s11292-022-09550-w
Kleeven, A. T. H., de Vries Robbé, M., Mulder, E. A., & Popma, A. (2020). Risk assessment in juvenile and young adult offenders: Predictive validity of the SAVRY and SAPROF-YV. https://pubmed.ncbi.nlm.nih.gov/32964720/
O’Shea, L. E., & Dickens, G. L. (2014). Short-Term Assessment of Risk and Treatability (START): Systematic review and meta-analysis. Psychological Assessment, 26, 990–1002. https://doi.org/10.1037/a0036794
Protective factors in violence risk assessment: Predictive validity of the SAPROF and HCR-20V3 (2020). International Journal of Forensic Mental Health, 19. https://doi.org/10.1080/14999013.2019.1643811
Sijtsema, J. J., Kretschmer, T., & van Os, T. (2015). The Structured Assessment of Violence Risk in Youth in a large community sample of young adult males and females: The TRAILS study. Psychological Assessment, 27, 669–677. https://doi.org/10.1037/a0038520
The predictive validity of three youth violence assessment instruments: The SAVRY, VRS-YV, and SAPROF-YV (2022). International Journal of Offender Therapy and Comparative Criminology, 66, 168–185. https://pubmed.ncbi.nlm.nih.gov/33167725/
Violence risk assessment instruments in forensic psychiatric populations: A systematic review and meta-analysis (2023). The Lancet Psychiatry, 10, 780–789. https://doi.org/10.1016/S2215-0366(23)00256-0
Webster, C. D., Martin, M. L., Brink, J., Nicholls, T. L., & Desmarais, S. L. (2006). Short-Term Assessment of Risk and Treatability: The case for a new structured professional judgment scheme. https://pubmed.ncbi.nlm.nih.gov/17171764/
World Health Organization (2026). Youth violence: Fact sheet. https://www.who.int/news-room/fact-sheets/detail/youth-violence