Psychology by BogdanovaThe Encyclopedia of Psychology for the Artificial Era

RESEARCH DOSSIER

Parenting Capacity, Child Custody, and Child Protection Evaluations

Abstract. This fundamental research monograph examines the scientific bases, procedures, evidentiary limits, and ethical responsibilities of forensic evaluations of parenting capacity, child custody, and child protection. It defines functional parenting capacity and distinguishes parenting-plan questions from maltreatment-related judicial decisions. A critical narrative synthesis integrates attachment research, developmental and ecological theories, parent–child observational methods, forensic interviews, psychometrics, child-protection risk prediction, family-violence research, intervention trials, and professional standards. Evidence indicates substantial constraints on translating group-level associations into individual legal forecasts, uneven validity of assessment tools, and important risks from confirmation bias, cultural inequity, and conflation of diagnosis with caregiving ability. The analysis establishes a defensible evaluation logic based on child-specific needs, corroborated behavior, alternative hypotheses, uncertainty, and proportionality of conclusions. Implications for artificial decision support are stated as methodological research questions rather than established clinical capabilities. This is a critical narrative research synthesis, not a systematic review, a prospective study, a custody recommendation, or an individual forensic opinion.

Keywords: parenting capacity; child custody evaluation; child protection evaluation; forensic psychology; parenting-plan evaluation; child maltreatment; child welfare; risk assessment; best interests of the child; parent–child observation; forensic interviewing; attachment theory; intimate partner violence; psychological assessment; family court; forensic ethics; Artificial Era

The Forensic Research Object and Its Decisions

Parenting capacity, child custody, and child protection evaluations are structured psychological inquiries undertaken for consequential legal decisions about caregiving, child safety, parental responsibilities, contact, placement, and services. Their common object is the relationship between observable caregiving behavior, a particular child's developmental needs, environmental support, and foreseeable harm. Parenting capacity is therefore a functional, contextual, and time-dependent construct. It concerns whether an adult can provide sufficiently safe, responsive, and developmentally appropriate care under conditions that actually exist or can realistically be supported. A person's diagnosis, personality profile, income, or family configuration does not itself establish that functional capacity. The American Psychological Association's current child-protection guidance makes the causal nexus between conduct and significant harm or imminent risk central to the evaluation American Psychological Association (2024) National Academies (2016).

Three distinct decisions structure the field. Parenting-plan evaluation usually concerns the allocation of decision-making responsibility, residence, and contact after family separation. Child-protection evaluation examines alleged maltreatment, safety, caregiving deficits, and the conditions for preservation or restoration of care. Focused parenting-capacity evaluation examines specified caregiving tasks, including supervision, protection, responsiveness, and use of services. These inquiries may overlap but rely on different referral questions, decision standards, information sources, and consequences. A properly bounded forensic opinion answers the question actually referred, while acknowledging that an evaluator supplies psychological evidence and the authorized decision-maker applies the law American Psychological Association (2022) Association of Family and Conciliation Courts (2022) American Psychological Association (2024).

The monograph synthesizes historical research, original empirical studies, observational measurement, practitioner surveys, professional guidelines, risk-tool reviews, intervention trials, and methodological criticism. Its method is a critical narrative research synthesis, not a prospectively registered systematic review. Its central questions are how caregiving competence can be observed, how child-specific risk is inferred, how error and bias enter adjudicative evidence, which interventions change outcomes, and how expert conclusions can remain proportionate to the evidence. The analysis is primarily informed by American forensic practice, with comparative research used only where the institutional context is explicit Budd et al. (2001) McNellan et al. (2022) Association of Family and Conciliation Courts (2022).

Parenting Capacity as a Functional and Relational Construct

A useful operational account separates caregiving opportunity, caregiving behavior, and caregiving outcome. Opportunity describes the resources and conditions that permit an adult to perform parenting tasks; behavior describes demonstrated supervision, protection, emotional availability, rule setting, health-related care, and adjustment to the child; outcome describes evidence of safety, distress, development, and relational functioning. The distinction matters because a short observation can show competent behavior without proving long-term stability, whereas an adverse outcome may arise from preexisting illness, outside trauma, or institutional disruption. Capacity claims must connect tasks to the child's developmental stage and assess whether needed supports are durable, accessible, and acceptable National Academies (2016) Budd et al. (2001).

Parenting is distributed across households, kinship networks, schools, health care, and social services. A caregiver may organize safe care through reliable partners and adaptations rather than personally performing every task. The evaluation therefore asks whether protective functions are accomplished, how responsibility is coordinated, what happens when arrangements fail, and whether protective resources are realistically available. This functional view reduces unjustified inferences from disability, poverty, employment arrangements, and nontraditional family structures. A balanced assessment records strengths, protective relationships, and the observed effects of services with the same rigor applied to deficiencies American Psychological Association (2024) National Academies (2016).

Historical Formation of Family-Court Psychological Expertise

Family-court psychology developed through the convergence of child-development research, custody adjudication, child welfare investigation, clinical assessment, and forensic standards. Earlier evaluations often emphasized global parental traits or diagnoses; later methodological work pressed for explicit referral questions, child-specific evidence, multiple information sources, and methods whose validity applies to the legal inference being made. The APA's revised custody guidelines of February 2022 and the AFCC's May 2022 parenting-plan guidelines reflect that shift toward clear role boundaries, methodological balance, informed practice, and transparent reasoning American Psychological Association (2022) Association of Family and Conciliation Courts (2022).

Empirical critique also shaped the field. Budd, Poindexter, Felix, and Naik-Polan analyzed 190 reports from a large urban juvenile court system and identified common weaknesses: single-session evaluations, limited collateral information, few behavioral methods, and insufficient description of actual caregiving. Their 2001 study is a historically important examination of practice quality rather than a nationally representative measurement of every subsequent evaluation. Surveys of custody professionals similarly reveal differences in test use, specialization, and handling of domestic violence. Such studies explain why professional guidance developed as an evolving framework rather than a validation certificate for any single practitioner Budd et al. (2001) Quinnell and Bow (2001) Bow and Boxer (2003).

Children's Rights, Safety, Stability, and Participation

The child's safety and developmental interests organize the forensic inquiry. Relevant interests include protection from violence and neglect, continuity of sufficiently safe caregiving, dependable relationships, education, health, disability accommodations, and opportunities to express views without coercion. A child's preference is an evidentiary datum whose meaning depends on age, understanding, experiences, loyalties, fear, and the conditions under which the preference was elicited. It cannot be treated as a plebiscite or disregarded because its content conflicts with an adult's account American Psychological Association (2022) Forslund et al. (2022).

Safety and continuity interact rather than occupying separate decision worlds. Abrupt removal from dangerous care may protect physical integrity while creating new relational and placement disruptions; leaving a child in an unsafe setting may preserve familiarity while maintaining exposure to harm. The evaluation should therefore compare realistic, time-specific options and describe the foreseeable burdens of each. Attachment research supports continuity of familiar nonabusive care and access to multiple supportive attachment figures, while cautioning against interpreting a one-time attachment classification as a court-ready ranking of caregivers Forslund et al. (2022) Forslund et al. (2021).

Theories of Attachment, Caregiving, and Development

Attachment theory explains how children organize proximity-seeking and expectations of care in response to patterns of protection and responsiveness. It supplies a developmental framework for continuity, relationships, and distress during separation. A 2022 international consensus statement stressed three propositions for court use: familiarity with nonabusive caregivers, continuity of adequate care, and the developmental value of networks of attachment relationships. These propositions should orient questions about caregiving histories and placement transitions, while the predictive value of individual attachment classifications in adjudication remains disputed Forslund et al. (2022).

Forslund, Hammarlund, and Granqvist distinguished attachment research as framework evidence from individual diagnostic evidence. Group-level associations among caregiving, attachment, and development cannot establish which parent in a specific dispute will provide safer care or produce a superior long-term outcome. An evaluator who substitutes a child's attachment label for observed caregiving skill commits a level-of-inference error. Attachment assessments may help plan support or understand behavior, but individualized legal conclusions need wider histories, observations, risk evidence, and causal analysis Forslund et al. (2021) Gridley et al. (2019).

Ecological, Transactional, and Family-System Models

Ecological models situate parental behavior within housing, employment, violence, service access, discrimination, neighborhood conditions, and support networks. Transactional models examine how children and caregivers mutually influence each other's behavior over time. Family-system approaches attend to alliances, communication, conflict, boundary management, and the coordination of care. These models prevent a narrow trait-based reading of parenting: a child's dysregulation may increase caregiving demands, and a caregiver's responses may subsequently amplify or reduce that dysregulation. Longitudinal data are needed to separate temporal order from reciprocal and confounded processes National Academies (2016).

The forensic implication is to specify a causal account that links a proposed risk factor to the child's actual experience. For example, unemployment may operate through resource deprivation, but its effect differs where stable housing, kinship support, or assistance is available. A mental-health symptom might impair supervision during acute episodes while leaving substantial competence between episodes; the relevant questions concern severity, functional consequences, treatment response, supports, and safety planning. Decision-makers need these pathways and their uncertainty, not mere inventories of adverse circumstances American Psychological Association (2024) Lane and Seltzer (2023).

Designing a Referral-Specific Assessment Protocol

Good forensic evaluation begins with a written decision question, explicit candidate explanations, and a data plan matched to the question. The evaluator identifies the child, caregivers, relevant placements, jurisdiction, time period, allegations, and information needed to distinguish competing hypotheses. The order or agreement should specify authority, records access, deadlines, reporting obligations, and the treatment of incomplete participation. Methods then follow the evidentiary task: observations for caregiving behavior, records for chronology, interviews for accounts and meaning, standardized measures for defined constructs, and collaterals for functioning across contexts Association of Family and Conciliation Courts (2022) American Psychological Association (2022).

An investigation of alleged neglect, for example, must distinguish unmet need, parental knowledge, available resources, and actual consequences. An assessment of a proposed shared-care schedule must ask about routine stability, school logistics, safety, developmental needs, parental cooperation where feasible, and prior caregiving arrangements. The protocol should identify potential disconfirming evidence before the first interview. This design reduces hindsight bias and the tendency to collect facts that merely support the appointing party's theory. Documentation should allow another expert to reconstruct the reasoning without agreeing with its conclusion American Psychological Association (2024) Association of Family and Conciliation Courts (2022).

Interviews with Parents and Other Caregivers

Parent interviews supply accounts of daily care, educational and medical routines, household composition, behavioral guidance, conflicts, life events, and the caregiver's understanding of the child's needs. They also reveal plans, explanations, and perceived obstacles. Because high-stakes litigation creates incentives toward impression management, self-report is necessarily one evidentiary channel among several. Parallel coverage of relevant topics across caregivers enhances procedural fairness while allowing individualized follow-up. Interviewers should document inconsistencies, reasonable explanations, language barriers, and opportunities for correction Quinnell and Bow (2001) Association of Family and Conciliation Courts (2022).

Structured questioning improves completeness but cannot produce validity by itself. Interview behavior may reflect anxiety, trauma, neurodivergence, culture, language proficiency, or mistrust of institutions. The evaluator should separate observations about communication style from claims about parental competence and avoid treating emotion displayed during litigation as a direct measure of caregiving. Where there is coercive control or violence, symmetrical interviewing methods may create asymmetric safety consequences. Disclosure and collateral contact therefore require explicit attention to confidentiality limits and retaliation risk American Psychological Association (2022) Saunders et al. (2013).

Interviewing Children and the Science of Suggestibility

Child interviews seek an age-appropriate account of experiences, routines, perceived safety, attachment figures, and wishes. The method must minimize suggestion and distinguish what the child personally experienced from what was later learned through adults or prior interviews. Investigative interviews concerning suspected maltreatment have a specialized evidentiary purpose; trained forensic interviewers should conduct them within coordinated procedures. Lamb, Orbach, Hershkowitz, Esplin, and Horowitz reviewed research on the NICHD protocol and described improvements in the quality and informativeness of investigative interviews when structured, open-ended, developmentally appropriate techniques were used Lamb et al. (2007) APSAC (2023).

Repeated interviews can increase opportunity for information while also introducing inconsistency, rehearsal, and suggestive contamination. In custody and protection work, a child may have spoken with caregivers, therapists, social workers, attorneys, and investigators before the evaluator becomes involved. The assessor must document who asked what, preserve distinctions between spontaneous narrative and prompted responses, and avoid drawing a truthfulness conclusion from confidence, fluency, emotional expression, or narrative detail alone. The most relevant evidence concerns the provenance, conditions, and consistency of statements within the full record APSAC (2023) Lamb et al. (2007).

Parent–Child Interaction and Observational Sampling

Direct observations can show how a caregiver notices cues, provides comfort, coordinates transitions, supports autonomy, and responds to frustration. They are valuable because they sample functional behavior rather than relying entirely on retrospective narrative. Yet a structured office observation is a limited situation with an observer present, unfamiliar tasks, and participants aware of legal consequences. A brief positive encounter cannot establish long-term protection, and a difficult encounter may reflect recent separation, illness, prior disruption, or the experimental context. Observation is best interpreted as an episode within a longitudinal and multisource account Gridley et al. (2019) American Psychological Association (2022).

Measurement quality requires a specified coding system, trained raters, clear construct definitions, documented interrater agreement, and evidence that scores mean what the evaluation claims. Gridley and colleagues' 2019 systematic review found weaknesses in the psychometric evidence for commonly used observational measures of parent–child interaction, especially generalization across the entire birth-to-five age span. This is a warning against moving from a validated research outcome measure to an unvalidated individual custody ranking. Structured tasks should be adapted to disability, language, sensory needs, and developmental age without pretending that modified conditions retain original norms Gridley et al. (2019) Association of Family and Conciliation Courts (2022).

Records, Collateral Sources, and Competing Accounts

Health and educational records, child welfare files, prior court orders, attendance information, contemporaneous communications, and observations by appropriately situated professionals can establish chronology and consequences unavailable from a single interview. A source's reliability depends on how information was obtained, whether it was observed directly, incentives to report, temporal proximity, and possible dependency on another source. Multiple reports repeating one original accusation constitute correlated evidence, not independent corroboration. Each contested fact should therefore carry an explicit provenance and a record of what would disconfirm it Association of Family and Conciliation Courts (2022) American Psychological Association (2024).

Collateral information also creates risks of selection bias. One caregiver may have greater access to schools and clinicians; another may work irregular hours or face language barriers. A professional may know a child only during crisis, and a grandparent may have strong interests in the outcome. The evaluator should provide reasonably comparable opportunities for the presentation of relevant material and identify missing records without silently treating absence as negative evidence. Evidence matrices can distinguish documented behavior, collateral assertion, evaluator observation, psychological inference, and unresolved allegation Budd et al. (2001) Association of Family and Conciliation Courts (2022).

Psychological Testing and the Problem of Construct Transfer

Standardized personality, symptom, cognitive, and behavioral instruments may clarify questions about functioning, response style, or needs for services. They do not automatically measure parenting competence or determine custody arrangements. A depression scale measures selected symptoms under its validation conditions; a finding of elevation supplies no direct estimate of the probability that a particular child will be harmed. Using a test requires checking the normative population, language, administration integrity, reliability, validity, incremental contribution, and whether the legally relevant construct is represented Quinnell and Bow (2001) American Psychological Association (2022).

Quinnell and Bow's national survey of 198 psychologists described testing as one source among several in custody evaluation practice. That survey records professional use and attitudes, not evidence that any resulting custody recommendation predicts better outcomes. The evaluator must differentiate a psychometric score from a forensic conclusion, report conditions that could alter interpretation, and use alternative evidence where the test is inappropriate. Computer-generated interpretive reports require the same scrutiny as human-generated interpretations, particularly when generic personality descriptors are presented as predictions about actual caregiving Quinnell and Bow (2001) Association of Family and Conciliation Courts (2022).

Response Validity, Defensive Presentation, and Informant Disagreement

Because custody and child-protection evaluations are high stakes, participants may emphasize positive qualities, minimize problems, or present strongly adverse accounts of others. The presence of a motive to influence the outcome does not establish deception, and inconsistency is not equivalent to malingering. Response-style indicators can be useful when their validity is documented for the specific test and setting, but culturally patterned self-presentation, defensive anxiety, misunderstanding, and genuine change are alternative explanations that must be considered American Psychological Association (2022) Association of Family and Conciliation Courts (2022).

Informant disagreement often contains clinically meaningful information. Parents may describe different periods, teachers different social settings, and children different relational experiences. Instead of mechanically averaging reports, the forensic analysis asks when and where each claim is plausible, which observations are independent, and how disagreements map onto competing hypotheses. The report should preserve unresolved variation and show how much each uncertain premise matters to the final recommendation. A finding that one informant's account is unsupported does not in itself validate the opposing account Budd et al. (2001) Association of Family and Conciliation Courts (2022).

Child Maltreatment, Neglect, and the Causal Nexus to Harm

Child-protection evaluations require a precise account of alleged harm and the caregiving mechanism by which it occurred or may occur. Physical violence, sexual abuse, emotional maltreatment, medical neglect, inadequate supervision, and exposure to domestic violence involve different evidence, thresholds, and alternative explanations. In neglect allegations, the assessment must distinguish absence of an essential resource from a caregiver's unwillingness or inability to use a resource that is actually available. The relevant psychological judgment concerns the effect on a specific child's safety, health, and development, as required by the referral and applicable law American Psychological Association (2024) National Academies (2016).

Causal claims must specify chronology, the child's preexisting vulnerabilities, exposure intensity, alternative causes, and realistic protective actions. An elevated risk factor is a prompt to investigate, not proof that maltreatment occurred. Conversely, absence of measured psychopathology cannot exclude dangerous conduct. Both harmful acts and omissions need corroboration appropriate to their nature. A defensible report records the level of confidence for each significant factual premise, the impact of uncertain allegations, and the mechanisms by which proposed arrangements would mitigate immediate and future harm American Psychological Association (2024) APSAC (2022).

Risk Prediction, Base Rates, and the Limits of Instruments

Risk instruments promise consistency but assess a specified outcome over a specified horizon in populations used for validation. The base rate of an outcome changes the proportion of positive classifications that are correct even when sensitivity and specificity remain constant. In low-base-rate settings, a seemingly informative tool can produce many false-positive designations. Practitioners must distinguish risk of an initial report, substantiation, placement disruption, serious injury, and recurrence; those outcomes are related yet conceptually distinct. A single score cannot answer all of them McNellan et al. (2022) Zhu et al. (2025).

A 2022 U.S. scoping review identified 25 studies of child-protection risk assessment instruments and judged the validity literature heterogeneous, dated, and limited with respect to demographic subgroups. A later meta-analysis encompassing 28 studies and 27 tools reported pooled discrimination of AUC 0.686; its abstract reported differing performance by instrument type. These findings describe average ranking performance and do not license precise individual probabilities without calibration, temporal and geographic transportability, subgroup analysis, or evidence of decision benefit. AUC is a measure of discrimination, not proof that a threshold produces beneficial child outcomes McNellan et al. (2022) Zhu et al. (2025).

Intimate Partner Violence, Coercive Control, and Safety Planning

Violence between adults may affect children directly through injury, coercive exposure, psychological distress, compromised caregiving, and postseparation threats. Assessment should distinguish patterns of coercive control, physical assault, intimidation, stalking, situational conflict, and reciprocal resistance. Treating all adult conflict as interchangeable can obscure unequal power, safety risk, and the reasons a parent limits contact. Violence-related evaluation must consider current exposure, escalation, access arrangements, protective orders, safe communication, and the likelihood that assessment procedures themselves could disclose dangerous information Saunders et al. (2013) American Psychological Association (2022).

In a 2013 study of 465 custody evaluators, Saunders, Tolman, and Faller found that evaluators' beliefs about intimate partner violence explained more variation in custody and visitation recommendations than measured demographic and knowledge variables. The study was observational and addressed professional decision processes, not the objective correctness of individual case outcomes. Its significance lies in showing how extra-evidentiary beliefs can enter apparently clinical conclusions. The report must therefore separate confirmed behavior, disputed allegations, risk mechanisms, and statutory obligations while making the role of evaluator assumptions inspectable Saunders et al. (2013) Bow and Boxer (2003).

Parent–Child Contact Resistance and Competing Explanations

Children may resist contact because of frightening experiences, abuse, disrupted caregiving, loyalty conflicts, anxiety, ordinary developmental preferences, inconsistent routines, pressure from an adult, or combinations of these conditions. An evaluator must investigate alternative pathways rather than assigning a single label from the intensity of the child's resistance. The APA's 2022 custody guidelines recognize parent–child contact problems as multifactorial and note the presence of both scientific literature and nonscientific accounts. Particular care is required where alleged alienating behavior competes with allegations of violence or serious neglect American Psychological Association (2022).

The task is to reconstruct the history of interactions, changes in the child's responses, the content and context of the child's statements, observed parental conduct, and the safety consequences of any proposed intervention. A parent who fails to encourage contact may be responding to verified risk, acting coercively, or both; the inference cannot be made from behavior alone. The language used in reports should describe observed acts and supported pathways, including the evidence for alternative explanations, rather than treating contested syndromes as validated individual diagnostic categories American Psychological Association (2022) Saunders et al. (2013).

Parental Mental Health, Substance Use, and Functional Assessment

Psychiatric symptoms, substance misuse, neurodevelopmental conditions, and cognitive limitations may be relevant where they affect the performance of caregiving tasks. Their relevance depends on functional impairment, time course, insight, recurrence, available treatment, practical support, and the child's needs. A diagnosis cannot substitute for analysis of whether a caregiver can supervise safely, respond to distress, sustain routines, or recognize danger. The same symptom may have very different consequences under stable treatment, family support, and service access American Psychological Association (2024) National Academies (2016).

For substance use, the inquiry addresses intoxication during caregiving, withdrawal, storage and exposure hazards, reliability, relapse prevention, recovery supports, and the consequences of actual incidents. For cognitive disability it addresses communication, task learning, support adaptations, and performance over time. Standardized assessment may inform one component, but direct task-based observation and realistic service planning are essential. Clinical expertise must also recognize that prediction across months or years is more uncertain than description of present functioning, especially when prospective support conditions are unstable American Psychological Association (2024) Budd et al. (2001).

Poverty, Cultural Context, Disability, and Equity

Poverty can coexist with unmet needs while reflecting institutional resource constraints rather than a disposition toward harmful care. Cultural practices influence communication, extended-family involvement, discipline norms, and the meaning assigned to children's autonomy. Disability can alter the manner in which parenting tasks are accomplished without eliminating competent care. The evaluator should specify which behavior poses a concrete risk, which interpretations rely on cultural assumptions, and which reasonable accommodations or supportive services could change the risk. A fair assessment requires equivalent scrutiny and access to evidence across caregivers Lane and Seltzer (2023) American Psychological Association (2024).

Lane and Seltzer's 2023 ethics analysis discusses racial inequities in suspicion, reporting, investigation, and child-welfare outcomes. That literature requires a distinction between an objectively dangerous condition and institutional pathways that differ by race, income, or service access. The evaluator should document how referrals originated, whether evidence is independent, how language assistance was provided, and whether comparison standards are appropriate. Equity is a property of the entire evidentiary process, including who is evaluated, which risks are measured, and whose protective resources are recognized Lane and Seltzer (2023) McNellan et al. (2022).

Intervention, Parenting Programs, and Causal Evidence

An assessment may identify remediable difficulties, but treatment efficacy must be evaluated by outcome. Parent training, coaching, home visitation, family support, and trauma-informed services target different mechanisms and populations. Evidence from randomized studies is stronger for some observed parenting behaviors than for low-frequency severe maltreatment outcomes. A 2023 meta-analysis by Backhaus and colleagues found reductions in reported physical and emotional violent parenting immediately after intervention, with smaller effects at follow-up; the evidence was frequently based on caregiver self-report and risk of bias affected several domains Backhaus et al. (2023).

An umbrella review published in 2024 found that no intervention category showed a clear, consistent, robust record of preventing or reducing child maltreatment across the broad literature, although positive findings occurred for specific approaches in specific settings. These conclusions can coexist with individual trials and meta-analyses reporting improvements: outcomes, comparators, populations, and follow-up intervals differ. Forensic recommendations should identify a concrete intervention target, evidence for that target, feasible implementation, and observable conditions under which safety would be reassessed Gautschi and Lätsch (2024) Backhaus et al. (2023).

Reunification, Alternative Care, and Longitudinal Outcomes

Decisions about reunification compare the child's present safety, developmental time, available family support, and the burdens of extended separation or placement changes. A return home can restore important relationships while reintroducing unresolved hazards. Continued alternative care can provide immediate protection while adding instability or relational loss. The evaluator's responsibility is to compare plausible trajectories from available evidence, document how services might alter those trajectories, and propose review conditions rather than treating either family preservation or removal as an automatic outcome Forslund et al. (2022) American Psychological Association (2024).

A systematic review of postreunification child-protection involvement identified child, family, placement, and service-level factors across 42 papers. The review emphasized individualized assessment and the complexity of needs after reunification; it did not validate a universal numerical decision rule. Follow-up evaluation should therefore monitor meaningful outcomes such as stable supervision, consistent school attendance, caregiver functioning, renewed incidents, child well-being, and placement continuity while recognizing that administrative re-referrals measure agency contact as well as underlying harm Tucker et al. (2025).

The Best-Interests Standard and the Limits of Expert Knowledge

The familiar best-interests standard commits a court to assessing a child's welfare while leaving the relative weight of competing values partly dependent on local law. Psychological evidence can clarify developmental consequences, exposure to harm, continuity of relationships, and the feasibility of support. It cannot independently determine how the law should trade these considerations against rights, procedural protections, and other normative commitments. Emery, Otto, and O'Donohue's 2005 critique placed this epistemic limitation at the center of custody assessment: a legally broad question can demand certainty far beyond the validated scope of psychological instruments. The important methodological correction is to divide an ultimate legal question into testable factual and functional subquestions and identify which judgments remain legal Emery et al. (2005) Association of Family and Conciliation Courts (2022).

An evaluator should not treat the vague breadth of a referral as permission to pronounce upon every aspect of family life. For example, the significance of prior caregiving arrangements depends on whether those arrangements were safe, freely chosen, and feasible under present conditions; evidence about historical roles can inform continuity but cannot alone resolve future risk. Likewise, a child's expressed wish helps the court understand the child's experience but requires evidence about developmental understanding, possible pressure, and realistic consequences. A transparent report states where psychology contributes and where the decision rests on a legal principle, contested value, or unresolved fact. This boundary protects the court's decision-making function and preserves the expert's scientific accountability American Psychological Association (2022) Emery et al. (2005).

Clinical Constructs and Forensic Decision Validity

Clinical diagnoses, intellectual functioning, personality measures, and symptom inventories were developed for questions that may differ substantially from legal questions about care. A sound clinical diagnosis may have utility in treatment planning, yet the inference from that diagnosis to unsafe parenting depends on additional assumptions. Emery and colleagues emphasized the gap between ordinary clinical validity and validity for a custody recommendation. That gap should be made visible in the case formulation: first describe the measured construct; next specify the observed caregiving behavior; then test whether the association has a plausible mechanism and sufficient evidence to affect a decision about a particular child Emery et al. (2005) Quinnell and Bow (2001).

The inference may also run in the opposite direction. A caregiver can provide effective and loving care despite a psychiatric diagnosis, while someone who scores within ordinary clinical ranges may engage in deliberate coercion, neglect, or violence. In high-conflict cases a defensive response style, elevated distress, or a dispute-related symptom can reflect transient circumstances rather than stable impairment. A report should indicate the incremental value of each test over records, interviews, and observed behavior. When a measure contributes only general characterization, it should be presented as such rather than converted into a numerical estimate of parenting adequacy American Psychological Association (2022) Emery et al. (2005).

Protective Caregiving in the Context of Domestic Abuse

Assessing a caregiver's protective conduct requires attention to what the person knew, could reasonably control, and could accomplish under prevailing threats. A nonoffending parent may face economic dependence, surveillance, violence, threats involving the children, or restricted access to housing and legal protection. The same apparent act—such as permitting contact or failing to report immediately—may reflect coercion, limited alternatives, misplaced judgment, or a combination. The evaluator's task is to reconstruct actionable choices at the relevant time rather than presume that an idealized protective response was available. Contemporary child protection guidance calls for linking parental action or inaction to actual or imminent harm American Psychological Association (2024) Saunders et al. (2013).

Protection should also be evaluated prospectively. A safety plan depends on genuine access to resources, its acceptability to the child and caregiver, the potential for retaliation, and agencies' capacity to implement promised protections. A proposed shared-care arrangement may require repeated direct exchanges that create opportunities for coercion; another arrangement may reduce exposure while preserving safe communication. Domestic-violence findings should therefore translate into specific pathway analysis, risk mitigation, and monitoring obligations. Evaluators must preserve uncertainty about disputed conduct while fully examining credible safety threats, rather than treating the absence of a criminal conviction as proof of safety American Psychological Association (2022) Association of Family and Conciliation Courts (2022).

Multidisciplinary Coordination and Avoidable Interview Burden

Child-protection matters may involve a pediatrician, therapist, school, social worker, police investigator, specialist forensic interviewer, guardian, and court evaluator. Their tasks differ: a pediatrician evaluates health, an investigator collects allegation-related evidence, a therapist supports treatment, and a psychologist appointed to the court addresses a defined psycholegal issue. The risk is that similar questions are asked repeatedly under varying assumptions and conditions, producing distress and source contamination. Effective coordination identifies the purpose of each contact, permissions for sharing records, the provenance of statements, and whether a recent high-quality interview already answers part of the referral APSAC (2023) American Psychological Association (2024).

Professional boundaries do not imply that perspectives should be isolated. A child's educational difficulties may be clarified by teachers, while medical records may specify unmet treatment needs and changes after intervention. The forensic evaluator can integrate these findings while preserving the original professional's scope and the distinction between direct observation and reported history. Specialized child interviews should be conducted under procedures designed for investigative reliability; therapeutic disclosure may have a different purpose and evidentiary quality. Agencies should favor one well-documented assessment over redundant procedures when the additional information is unlikely to change the decision Lamb et al. (2007) APSAC (2022).

Evidence Provenance, Independence, and the Problem of Repetition

High-stakes case files are frequently dense but evidentiary density is not the same as evidentiary independence. A school report may quote a child's parent, a child-welfare summary may reproduce the school report, and an expert may cite both as separate corroboration. The original information still has a single source, and repeated transmission can create an illusion of convergence. A useful evidence table therefore records the originating person, exact observation, date, method of acquisition, subsequent transformations, and known incentives or limitations. This is particularly important when retrospective statements become more definite across successive summaries Budd et al. (2001) APSAC (2023).

Contradictory records need structured reconciliation. Different dates, partial perspectives, genuine change, ambiguous terminology, or reporting errors can all explain discrepancies. The assessor should specify which claims are independently established, which are supported only by interested informants, and which remain unsettled. Digital material adds versioning and authenticity questions: screenshots can remove previous messages, video clips can omit surrounding events, and transcriptions can misstate speech. The evidentiary chain should be preserved and uncertainty carried into the report's recommendations. Expert confidence is justified by the quality and independence of evidence, not by the number of pages assembled Association of Family and Conciliation Courts (2022) American Psychological Association (2024).

Statistical Thresholds and the Costs of Wrong Decisions

Risk scoring involves a decision threshold as well as a prediction model. Consider a tool that labels families above a fixed score as high risk. Its consequences depend on how common the target outcome is, which harms matter most, and the costs of classifying families incorrectly. A threshold chosen to identify more future adverse events may increase false alarms and unnecessary intrusive investigations; a more restrictive threshold may fail to identify families in immediate danger. Statistical discrimination, calibration, decision curves, and practical implementation therefore answer different questions. The scoping and meta-analytic literatures do not establish a universal threshold for child protection decisions McNellan et al. (2022) Zhu et al. (2025).

The ethical stakes are unusually consequential because an incorrect positive inference may cause a serious disruption of family relationships, while a negative inference can leave a child in danger. The alternatives are often more numerous than removal versus no action: tailored assistance, supported supervision, kinship care, monitored contact, and review can change the distribution of harms. Research should evaluate full decision pathways rather than a model's AUC in isolation. Valid prediction also requires subgroup calibration, transparent handling of missing variables, and longitudinal tracking of both safety outcomes and intervention burdens Zhu et al. (2025) National Academies (2016).

Caregiving Change, Temporal Validity, and Reassessment

A capacity opinion is a statement about functioning under specified conditions and during a defined period. Parents may improve through sustained services, relapse during crisis, form more supportive networks, or confront new disability-related demands. The time needed for meaningful change also differs across safety mechanisms. An intervention might quickly improve scheduling or access to childcare while entrenched violence or a pattern of repeated harmful conduct requires longer assessment and effective risk control. Reassessment must therefore target behavior and consequences rather than presume that completion of sessions proves competence or that an earlier assessment remains permanently valid Backhaus et al. (2023) American Psychological Association (2024).

Follow-up measures should be defined before evaluation of improvement. Examples include consistently safe supervision, verified use of supports, independent reports from the child's daily environments, and sustained functioning across ordinary stressors. Child outcomes should also be tracked where valid and feasible; a short-term reduction in parent-reported harsh behavior cannot substitute for evidence concerning severe harm. Studies of postreunification involvement show that recurrent agency contact depends on multiple child, family, and system factors, so it is an imperfect proxy for abuse or successful care. The reporting interval and changing service environment must be explicit Tucker et al. (2025) Gautschi and Lätsch (2024).

Scientific Admissibility and the Reproducibility of Expert Opinions

Scientific admissibility concerns the relevance and reliability of reasoning offered to a legal decision-maker. A professional credential, a familiar psychological test, and an elaborate report do not by themselves establish that the specific causal claim has been validated. Forslund and colleagues' distinction between attachment as framework evidence and attachment as purported individual diagnostic evidence illustrates the issue. A developmental principle can be well established while the proposed instrument for ranking two caregivers lacks evidence of predictive validity in litigating families. The evaluator should explicitly identify this inferential distance Forslund et al. (2021) Emery et al. (2005).

Reproducibility in forensic work is more than repeating a questionnaire. It requires clear referral questions, accessible data provenance within lawful confidentiality limits, recorded observation rules, transparent methods for evaluating competing explanations, and a traceable bridge from findings to recommendations. Expert reports could be audited for whether independent reviewers identify the same facts, interpret measurement limitations similarly, and distinguish legal from scientific claims. Research could then compare whether more transparent procedures reduce disagreement or improve long-term child outcomes, while acknowledging that no single future outcome captures the full legal conception of best interests Association of Family and Conciliation Courts (2022) McNellan et al. (2022).

Children's Developmental Stage and Individual Needs

Parenting capacity exists in relation to a particular child at a particular stage of development. Infants depend on prompt protection and co-regulation; preschool children need predictable supervision, communication, and emotional support; school-age children require education and growing participation; adolescents need safety alongside expanding autonomy and privacy. Disability, chronic illness, prior trauma, and developmental variation may change these demands. The same care arrangement can function differently as needs and circumstances change, which makes age-independent declarations about an ideal contact schedule scientifically insufficient National Academies (2016) Forslund et al. (2022).

The assessor should specify the child's functional needs and what evidence shows that each caregiver and setting can meet them. Relevant questions include medication management, communication aids, school transitions, behavioral support, emotional recovery after contact, and relationships with siblings or other caregivers. Child reports must be interpreted developmentally while retaining their evidentiary importance. Normative developmental findings help identify questions; they cannot replace examination of the child's real daily experiences, risks, and protective relationships American Psychological Association (2022) American Psychological Association (2024).

Comparing Realistic Parenting Arrangements

Custody disputes often invite binary judgments about which adult is the better parent. A scientifically organized evaluation compares actual feasible arrangements: primary residence with scheduled contact, shared residence under specified conditions, supported transitions, supervised contact, kinship arrangements, or temporary protection with review. The comparative unit should include the child, the caregiving network, logistics, protection measures, and realistic resources. An arrangement that looks favorable abstractly may fail when travel, work, school, disability accommodations, or current danger are included in the model Association of Family and Conciliation Courts (2022) American Psychological Association (2022).

Evidence concerning joint physical custody illustrates why average outcomes cannot dictate an individual schedule. A 2026 scoping review of early attachment and joint physical custody identified only ten studies explicitly grounded in attachment theory. That narrow evidence base is relevant to the limits of categorical age-specific prescriptions. Recommendations should describe assumptions, feasible alternatives, monitoring criteria, and the conditions that would warrant revision; they should not present an untested universal schedule as a psychological fact Funkquist Sköld et al. (2026) Forslund et al. (2021).

The Empirical Quality of Existing Forensic Evaluation Practice

Studies examining completed reports and practitioners' views reveal several recurring tensions. Budd and colleagues' 190-report analysis found many evaluations insufficiently focused on actual parenting behavior. Quinnell and Bow's survey of 198 psychologists documented changes in testing practices, yet professional preference provides no validation of prediction. Bow and Boxer's survey of 115 evaluators found both practices that aligned with guidelines and underuse of specialized domestic-violence assessment tools. These studies address different samples and periods; their results should not be pooled as though they estimated a single prevalence of evaluation error Budd et al. (2001) Quinnell and Bow (2001) Bow and Boxer (2003).

Safer and more informative practice requires a separation of process quality, scientific validity, and outcome accuracy. A report may describe all interviews carefully while using poorly validated tests. It may employ validated tests while omitting the child's actual experiences. It may reach a plausible recommendation while failing to specify why that recommendation follows from the evidence. A comprehensive audit therefore reviews referral fit, data provenance, interpretive validity, response to contrary evidence, cultural competence, transparency, and the possibility of evaluating long-term outcomes Association of Family and Conciliation Courts (2022) American Psychological Association (2024).

Evidence Hierarchies, Causation, and Uncertainty

The evidentiary strength of a finding depends on the question. Court files can establish documented events but may preserve earlier errors; interviews capture accounts but are vulnerable to memory and incentive effects; direct observation captures behavior under limited conditions; validated instruments measure specified constructs; randomized trials can identify effects of an intervention in defined populations. No single method occupies the top of a universal hierarchy. Strong assessment triangulates methods whose weaknesses are sufficiently independent and explains remaining uncertainty Budd et al. (2001) Gridley et al. (2019) McNellan et al. (2022).

Risk prediction must also be distinguished from causal explanation. A prior report may forecast renewed agency contact because it predicts both family difficulty and agency attention; it may have limited specificity for actual serious harm. Correlations between symptoms and adverse child outcomes can reflect shared stress or reverse causation. Intervention efficacy indicates the effect of assigning a program, not proof that every individual who completes it can safely provide all forms of care. A careful report names alternative hypotheses and calibrates language to the strength of the available evidence Zhu et al. (2025) Backhaus et al. (2023).

Quantitative reporting should preserve denominators, effect measures, and confidence intervals. For example, Backhaus et al. reported an immediate standardized effect of d = −0.46 (95% CI −0.59 to −0.33) on physical and emotional violent parenting outcomes, while the pooled effects were smaller at later follow-up. These estimates summarize a selected set of randomized studies and do not establish individual safety guarantees. Similarly, AUC 0.686 for risk tools is an aggregate discriminatory estimate, not the probability that one identified parent will maltreat a child Backhaus et al. (2023) Zhu et al. (2025).

Bias, Adversarial Contexts, and Procedural Fairness

Adversarial referral conditions increase the salience of confirmation bias, anchoring to initial allegations, allegiance effects, and inconsistent scrutiny of parties. The task is to build a decision process that makes these errors harder to conceal. Evaluators can state competing hypotheses before gathering evidence, review source independence, identify missing material, and separate observed events from interpretive judgments. Equal consideration does not require treating objectively unequal risks as equivalent; it requires applying relevant standards consistently while accounting for case-specific safety and procedural conditions American Psychological Association (2022) Saunders et al. (2013).

Procedural fairness also concerns notice of the evaluator's role, informed consent or legally appropriate authorization, confidentiality limitations, interpreters, participation opportunities, record preservation, and avoidance of improper one-sided contacts. The AFCC's 2022 guidelines address competence, relationships, communication, recordkeeping, data gathering, reporting, and accountability. Where a safety concern requires restricted disclosure, the restriction should be lawful and explained without sacrificing the accuracy of evidence handling. The evaluator should remain identifiable as an independent scientific participant rather than an advocate retained to confirm a party's desired result Association of Family and Conciliation Courts (2022) American Psychological Association (2022).

Forensic Reports, Testimony, and the Logic of Recommendations

A complete forensic report should state its mandate, evaluated questions, participants, methods, dates of contact, records, limitations, factual findings, disputed claims, and the reasoning that connects evidence to each opinion. It should identify the limits of psychological expertise and avoid assigning a statutory outcome merely by adopting legal language. Reports benefit from explicit distinctions among observed behavior, reported behavior, clinical interpretation, risk hypothesis, and preferred policy or legal outcome. This architecture makes a recommendation inspectable by the court and challengeable by the parties Association of Family and Conciliation Courts (2022) American Psychological Association (2022).

Recommendations should correspond to supported mechanisms: increased supervision for demonstrated safety failures, adapted services for observable skill deficits, protective communication arrangements where coercion is substantiated, or focused reassessment where data are inconclusive. Claims of certainty about future parenting or child trajectories exceed the scope of available research in many cases. When the available evidence supports several reasonable options, the report should explain the conditional tradeoffs and the threshold findings that would distinguish them. Transparent uncertainty is particularly important because evaluations can influence contact, placement, and family relationships for years American Psychological Association (2024) McNellan et al. (2022).

Ethics, Confidentiality, and Child-Centered Safeguards

Forensic evaluation requires competence in developmental psychology, assessment, maltreatment, family violence, legal procedure, cultural context, and the ethics of third-party decision support. Role boundaries matter because therapeutic relationships and court-appointed assessment answer different questions under different expectations of confidentiality. Evaluators must explain their role and reporting pathways, protect records, avoid unnecessary disclosure, and take safeguarding and mandated-reporting obligations seriously within the applicable jurisdiction American Psychological Association (2022) American Psychological Association (2024).

Safeguards should minimize assessment-related harm to the child. Repetitive questioning, unnecessary reexposure to alleged perpetrators, prolonged high-conflict observation, and poorly coordinated professional interviews can impose burdens and distort evidence. The evaluation should use existing high-quality records when appropriate and coordinate specialized forensic child interviews with trained professionals. Ethical review includes asking whether another measurement materially improves the legal question relative to its burden and whether a less intrusive source of equivalent information exists APSAC (2023) Association of Family and Conciliation Courts (2022).

Digital Evidence and Artificial Decision Support

Digital records increasingly enter custody and protection cases through messages, video calls, parenting applications, school portals, and social-media records. Such material may clarify timing and behavior, but selection, missing context, editing, and authenticity remain concerns. Analysts should preserve provenance, distinguish original messages from screenshots or summaries, and assess whether algorithmic sorting or selective exports omitted material. The psychological meaning of communication often depends on the relationship history, safety conditions, and context in which it was produced Association of Family and Conciliation Courts (2022).

Artificial systems may assist with document organization, transcription, chronology, language support, and detection of missing data. Their use raises research questions about transcription error, biased classification, hidden model assumptions, privacy, differential accuracy across populations, and misleading numerical precision. Any risk classification or automated recommendation requires task-specific validation, human accountability, and transparent error assessment. The Artificial Era creates opportunities for reproducible evidence management while increasing the importance of keeping causal interpretation and legal responsibility explicit. These are methodological implications and research hypotheses, not findings that automated custody judgments have been validated McNellan et al. (2022) Zhu et al. (2025).

Research Priorities and Unresolved Scientific Problems

The field needs prospective studies connecting evaluation procedures to meaningful child outcomes, rather than merely documenting recommendations. Multisite cohorts could assess the stability of observed parenting functions, differences among collateral and direct observation data, trajectories after different legal arrangements, and rates of preventable harm. Prediction studies should preregister outcomes and time horizons, report calibration as well as discrimination, test performance across demographic and disability groups, and evaluate the consequences of threshold decisions. Randomized trials can test service components where ethical and feasible McNellan et al. (2022) Tucker et al. (2025).

Additional priorities include measuring coercive control without endangering survivors, developing culturally valid observation protocols, understanding sibling and kinship networks, evaluating the child's participation without suggestion, and comparing support-first responses with other child-protection pathways. Attachment research requires clear constraints on the translation of group findings into individualized legal evidence. Intervention studies need independently observed safety outcomes and longer follow-up, because changes in reported parenting practices can occur without demonstrated reductions in serious maltreatment or out-of-home placement Forslund et al. (2022) Gautschi and Lätsch (2024) Backhaus et al. (2023).

An especially urgent methodological goal is to distinguish improvement in the apparent quality of an expert report from improvement in the accuracy and equity of the legal decision it informs. Transparent reasoning, calibrated uncertainty, and procedural fairness are necessary process safeguards; long-term child outcomes, reversibility of error, and unintended harm must also be measured. Evaluation science will mature when it can compare competing procedures on those outcomes and specify where added complexity improves decisions rather than merely increasing the appearance of expertise Budd et al. (2001) Association of Family and Conciliation Courts (2022).

Conclusion

Parenting capacity, custody, and child-protection evaluations are forensic applications of developmental, clinical, social, and methodological psychology to decisions that can alter children's lives. The strongest scientific approach defines the referral question, observes functional care, establishes chronology, tests competing explanations, and analyzes the child's specific needs. It integrates relevant collateral material, validated assessment where fit for purpose, and knowledge of safety and developmental mechanisms. It makes explicit how parental behavior, support networks, and environmental conditions interact to shape feasible decisions American Psychological Association (2022) American Psychological Association (2024) National Academies (2016).

The evidence also establishes important limits. Attachment research supplies developmental principles but rarely a valid individual caregiver ranking; risk instruments offer uncertain prediction whose quality depends on calibration and population; intervention effects vary by outcome and time; and evaluator expectations can influence recommendations. Scientific responsibility requires accurately scoped conclusions, transparency about alternatives, protection from assessment-related harm, and a commitment to revising judgments when new evidence appears. The aim of the field is defensible reasoning in service of child safety and development, with empirical accountability for the consequences of its recommendations Forslund et al. (2021) McNellan et al. (2022) Saunders et al. (2013) Backhaus et al. (2023).

References

  • American Psychological Association. (2022). APA guidelines for child custody evaluations in family law proceedings. https://www.apa.org/about/policy/child-custody-evaluations.pdf
  • American Psychological Association. (2024). APA guidelines for psychological evaluations in child protection matters. https://www.apa.org/about/policy/psychological-evaluations-child-protection-matters
  • Association of Family and Conciliation Courts. (2022). Guidelines for parenting plan evaluations in family law cases. https://www.afccnet.org/Resource-Center/Practice-Guidelines
  • Association of Professionals Solving the Abuse of Children. (2022). Forensic mental health evaluations when child maltreatment is at issue [Practice guidelines]. https://apsac.org/practice-guidelines/
  • Association of Professionals Solving the Abuse of Children. (2023). Forensic interviewing of children [Practice guidelines]. https://apsac.org/wp-content/uploads/2025/08/FI-Guidelines-Updated-V2.pdf
  • Backhaus, S., Leijten, P., Jochim, J., Melendez-Torres, G. J., & Gardner, F. (2023). Effects over time of parenting interventions to reduce physical and emotional violence against children: A systematic review and meta-analysis. EClinicalMedicine, 60, 102003. https://doi.org/10.1016/j.eclinm.2023.102003
  • Bow, J. N., & Boxer, P. (2003). Assessing allegations of domestic violence in child custody evaluations. Journal of Interpersonal Violence, 18(12), 1394–1410. https://doi.org/10.1177/0886260503258031
  • Budd, K. S., Poindexter, L. M., Felix, E. D., & Naik-Polan, A. T. (2001). Clinical assessment of parents in child protection cases: An empirical analysis. Law and Human Behavior, 25(1), 93–108. https://doi.org/10.1023/A:1005696026973
  • Emery, R. E., Otto, R. K., & O'Donohue, W. T. (2005). A critical assessment of child custody evaluations: Limited science and a flawed system. Psychological Science in the Public Interest, 6(1), 1–29. https://doi.org/10.1111/j.1529-1006.2005.00020.x
  • Forslund, T., Hammarlund, M., & Granqvist, P. (2021). Admissibility of attachment theory, research and assessments in child custody decision-making? Yes and No! New Directions for Child and Adolescent Development, 2021(180), 125–140. https://doi.org/10.1002/cad.20447
  • Forslund, T., Granqvist, P., van IJzendoorn, M. H., Sagi-Schwartz, A., Glaser, D., Steele, M., Hammarlund, M., Schuengel, C., Bakermans-Kranenburg, M. J., Steele, H., Shaver, P. R., Lux, U., Simmonds, J., Jacobvitz, D., Groh, A. M., Bernard, K., Cyr, C., Hazen, N. L., Foster, S., ... Duschinsky, R. (2022). Attachment goes to court: Child protection and custody issues. Attachment & Human Development, 24(1), 1–52. https://doi.org/10.1080/14616734.2020.1840762
  • Funkquist Sköld, L., Alfredsson, E., & Psouni, E. (2026). Joint physical custody and attachment in early childhood: A scoping review of the research landscape. Journal of Family Psychology, 40(2), 173–184. https://doi.org/10.1037/fam0001413
  • Gautschi, J., & Lätsch, D. (2024). The effectiveness of interventions to prevent and reduce child maltreatment in high-income countries: An umbrella review. Child Abuse & Neglect, 153, 106845. https://doi.org/10.1016/j.chiabu.2024.106845
  • Gridley, N., Blower, S., Dunn, A., Bywater, T., Whittaker, K., & Bryant, M. (2019). Psychometric properties of parent-child (0–5 years) interaction outcome measures as used in randomized controlled trials of parent programs: A systematic review. Clinical Child and Family Psychology Review, 22(2), 253–271. https://doi.org/10.1007/s10567-019-00275-3
  • Lamb, M. E., Orbach, Y., Hershkowitz, I., Esplin, P. W., & Horowitz, D. (2007). A structured forensic interview protocol improves the quality and informativeness of investigative interviews with children: A review of research using the NICHD Investigative Interview Protocol. Child Abuse & Neglect, 31(11–12), 1201–1231. https://doi.org/10.1016/j.chiabu.2007.03.021
  • Lane, W. G., & Seltzer, R. R. (2023). How should clinicians and health care organizations promote equity in child abuse and neglect suspicion, evaluation, and reporting? AMA Journal of Ethics, 25(2), E133–E140. https://doi.org/10.1001/amajethics.2023.133
  • McNellan, C. R., Gibbs, D. J., Knobel, A. S., & Putnam-Hornstein, E. (2022). The evidence base for risk assessment tools used in U.S. child protection investigations: A systematic scoping review. Child Abuse & Neglect, 134, 105887. https://doi.org/10.1016/j.chiabu.2022.105887
  • National Academies of Sciences, Engineering, and Medicine. (2016). Parenting matters: Supporting parents of children ages 0–8. The National Academies Press. https://doi.org/10.17226/21868
  • Quinnell, F. A., & Bow, J. N. (2001). Psychological tests used in child custody evaluations. Behavioral Sciences & the Law, 19(4), 491–501. https://doi.org/10.1002/bsl.452
  • Saunders, D. G., Tolman, R. M., & Faller, K. C. (2013). Factors associated with child custody evaluators' recommendations in cases of intimate partner violence. Journal of Family Psychology, 27(3), 473–483. https://doi.org/10.1037/a0032164
  • Tucker, E., O'Donnell, M., Krakouer, J., & Octoman, O. (2025). Reinvolvement after returning home: A systematic review of the factors associated with post-reunification child protection involvement. Child Abuse & Neglect, 167, 107515. https://doi.org/10.1016/j.chiabu.2025.107515
  • Zhu, N., Pan, X., & Zhao, F. (2025). Assessing the predictive validity of risk assessment tools in child health and well-being: A meta-analysis. Children, 12(4), 478. https://doi.org/10.3390/children12040478