Psychology by BogdanovaThe Encyclopedia of Psychology for the Artificial Era

RESEARCH DOSSIER

Interparental Conflict and Child Adjustment

Abstract. Interparental Conflict and Child Adjustment examines how the form, frequency, meaning, and resolution of conflict between caregivers relate to children's emotional, behavioral, physiological, social, and academic development. This critical narrative monograph integrates the cognitive-contextual framework, emotional security theory, family spillover and triangulation models, developmental psychopathology, and stress-regulation research. Its evidence base includes original observational and experimental studies, psychometric validation, genetically sensitive research, large meta-analyses, and intervention trials. The synthesis distinguishes direct exposure from child appraisals, parenting pathways, and the specific risks associated with intimate partner violence. Evidence consistently documents associations while showing that causal mediation, developmental heterogeneity, and intervention transportability remain open empirical questions. The concluding agenda emphasizes multidimensional measurement, prospective designs, ethical safeguards, and testing mechanisms across diverse family contexts.

Keywords: interparental conflict; child adjustment; emotional security theory; cognitive-contextual framework; children's perceptions of interparental conflict; child self-blame; family triangulation; feeling caught between parents; parenting spillover; internalizing symptoms; externalizing behavior; developmental psychopathology; coparenting; postdivorce family adjustment; intimate partner violence; sleep regulation; family intervention

Scientific Definition, Scope, and Units of Analysis

Interparental conflict refers to disagreement, hostility, withdrawal, or antagonistic interaction between adults occupying parental or caregiving roles in a child's family environment. Its scientific significance depends on the form, intensity, frequency, duration, content, and resolution of the interaction as well as the child's actual exposure and interpretation. Child adjustment denotes developmentally situated functioning across emotional symptoms, behavioral regulation, relationships, learning, sleep, and adaptive competence. A single composite score cannot represent every aspect of that functioning. The research object of this monograph is the relation between the organization of parental conflict and children's developing regulatory systems, with explicit attention to plausible mechanisms and the limits of causal attribution. The major reviews identify variation among conflict dimensions rather than treating parental discord as one uniform exposure van Eldik et al. (2020) Harold & Sellers (2018).

The relevant unit of exposure may be an individual episode, a recurring interaction pattern, a developmental history, or an organized family context. Episodes vary from resolved disagreements to hostile escalation, silent withdrawal, unresolved tension, and violence. A child's location, perceptual access, prior experience, and perceived personal implication determine whether nominally similar events carry comparable meaning. Family-level measures derived from one parent may conceal differences among children's experiences. A valid design therefore specifies whose conflict was measured, which relationship it concerns, what the child encountered, and whether the outcome reflects current distress or sustained difficulties. The cognitive-contextual framework made the child's appraisal central to these distinctions, while later meta-analytic work expanded the dimensional description Grych & Fincham (1990) van Eldik et al. (2020).

The phrase child adjustment requires careful operationalization. Internalizing problems include anxiety, depressive symptoms, withdrawal, and related distress; externalizing problems include aggression, rule-breaking, and difficulties with behavioral control. Adjustment also encompasses age-appropriate social participation, academic engagement, physical regulation, and perceived security. These domains can vary independently, and observers may see different parts of a child's functioning. A child might maintain school performance while experiencing persistent anticipatory anxiety at home, or exhibit irritability without a clinically established disorder. A comprehensive research design requires multiple indicators measured over time and prohibits extrapolating a diagnostic label from an elevated questionnaire score alone. The broader review literature supports assessment across psychopathology and adaptive developmental outcomes Harold & Sellers (2018) El-Sheikh et al. (2007).

This work is a critical narrative synthesis of experimental paradigms, longitudinal studies, cross-sectional surveys, psychometric validation, genetically sensitive designs, meta-analyses, and randomized intervention trials. The evidence is appraised according to temporal ordering, informant independence, construct validity, sampling, confounding, and the distinction between observed effects and explanatory models. Specific figures are reported only where a reviewed abstract or article provides them. The scientific objective is to explain which features of interparental conflict predict adjustment, through which tested and proposed pathways, under what developmental and contextual conditions, and what forms of intervention have demonstrated benefits. It includes violence as a distinct high-risk exposure requiring safeguarding and separate interpretation Kitzmann et al. (2003) van Dijk et al. (2020).

Historical Development of the Research Problem

Early family research often studied marital distress, separation, and divorce as broad household characteristics associated with children's difficulties. These descriptors helped establish family-level correlates of development, but their categories mixed structural events with ongoing relationship processes. Observational studies subsequently examined how children respond to specific adult emotional exchanges, including anger and subsequent reconciliation. A 1989 experiment using standardized adult interaction episodes, described by Cummings and colleagues, demonstrated a methodological transition toward children's observable responses to conflict histories and changes within an interaction. That line of work supplied behavioral questions that could be examined independently of parents' global satisfaction ratings Cummings et al. (1989) Harold & Sellers (2018).

Grych and Fincham's 1990 cognitive-contextual framework organized the child's interpretation of conflict into an explicit theoretical account. Attention shifted to perceived threat, responsibility, expectations about resolution, and available coping resources. The 1992 Children's Perception of Interparental Conflict Scale subsequently operationalized conflict properties, threat, and self-blame using reports from children rather than relying only on adult descriptions. Its initial validation involved 222 children aged nine to twelve, with cross-validation in 144 children of similar ages. These achievements gave researchers a way to distinguish environmental events from the psychological meanings children construct when exposed to those events Grych & Fincham (1990) Grych et al. (1992).

Davies and Cummings' emotional security hypothesis, published in 1994, advanced a complementary developmental account: the safety and predictability of the relationship between caregivers organize children's emotional reactions, behavioral involvement, and internal representations. The theory broadened the temporal question from immediate distress to how repeated experiences may influence subsequent regulation and adjustment. In parallel, Erel and Burman's 1995 meta-analysis of 68 studies examined the links between marital and parent-child relations and found support for a general spillover pattern. Together these programs distinguished appraisal, security, parenting, and broader family functioning while raising questions about their causal connections Davies & Cummings (1994) Erel & Burman (1995).

The field then expanded through multi-informant longitudinal designs, physiological recordings, sleep measurement, genetically informed comparisons, and structured intervention trials. The 2018 review by Harold and Sellers argued for examining the entire continuum of conflict expression, extending from withdrawal through aggression, instead of restricting inquiry to divorce or severe violence. The 2020 synthesis by van Eldik and colleagues compared six dimensions of the interparental relationship across large bodies of empirical evidence. Contemporary work also isolates triangulation and postseparation parenting pathways with meta-analytic methods. Scientific history is consequently best understood as increasing differentiation of exposure, psychological process, and outcome rather than a succession of mutually exclusive theories Harold & Sellers (2018) van Eldik et al. (2020) Schrodt (2026a).

Constructive Disagreement, Destructive Conflict, and Violence

Disagreement is an ordinary feature of relationships involving competing needs, goals, and interpretations. Research becomes informative when it distinguishes how disagreements are enacted and resolved. Constructive episodes may contain visible difference followed by explanation, mutual accommodation, or credible repair; destructive episodes may involve contempt, escalating hostility, unresolved anger, coercive control, or threat. Parents' conflict frequency alone cannot convey these qualities. Destructive expressions can be consequential even when adults subsequently report acceptable relationship satisfaction, whereas an observable disagreement accompanied by resolution may communicate that disagreement can be managed. The evidence base accordingly differentiates hostility, disengagement, unconstructive conflict, and child-related content van Eldik et al. (2020) Harold & Sellers (2018).

Child-related disagreements deserve separate attention because the child may infer direct responsibility for the dispute, become uncertain about parental alliances, or anticipate punishment and abandonment. The cognitive-contextual model predicts that the same expressed intensity can produce distinct responses depending on whether the child interprets the argument as threatening and personally caused. In experimental vignette research reported by Grych and Fincham, children reacted differently when conflict concerned them, illustrating the value of experimentally varying content while holding aspects of the scenario constant. Such tasks measure immediate reported reactions within a constructed setting; they do not establish the cumulative effects of real households by themselves Grych & Fincham (1993) Grych & Fincham (1990).

Coercive control, physical assaults, and intimate partner violence require an analytically and ethically distinct framework. Their defining characteristics include intimidation, danger, and asymmetrical power, with potential exposure to injury and ongoing fear. They should never be collapsed into a neutral notion of mutual disagreement or interpreted as interchangeable with ordinary couple communication. Kitzmann and colleagues' meta-analysis of 118 studies identified associations between children's exposure to interparental violence and psychosocial difficulties, while also documenting variations in study design and outcomes. In this exposure domain, safety assessment, survivor autonomy, and protection of children have priority over strategies presuming symmetrical responsibility or safe joint conflict discussion Kitzmann et al. (2003) World Health Organization (2026).

Disengagement and unresolved silence are also relevant even when overt aggression is absent. A child can perceive tension through changes in responsiveness, routine, facial expression, or household coordination. Nevertheless, brief adult withdrawal, solitary emotion regulation, and systematic coercive silent treatment are different processes; observational coding must distinguish them rather than label every pause harmful. The child's exposure should be assessed together with resolution signals and prior predictability. A design comparing conflict forms should measure adult behavior, child interpretations, and later adjustment separately, since correlations from the same reporter can reflect shared response tendencies. The 2020 dimensional synthesis provides the most appropriate starting point for testing distinctions among destructive expressions van Eldik et al. (2020) Harold & Sellers (2018).

Developmental Systems and Family Context

Children develop within interacting biological, relational, and institutional systems. The influence of an interparental episode depends on earlier experiences, developmental capacities, caregiver behavior, social resources, and wider conditions that shape household stress. The same child can show different responses across transitions in family structure, school demands, and relationships with siblings or peers. A developmental interpretation therefore requires an exposure history and an outcome trajectory rather than assuming that one reaction represents a stable trait. Prospective research can begin to address whether changes in family conflict precede changes in child functioning, provided that baseline difficulties and family context are measured with sufficient care Harold & Sellers (2018) van Dijk et al. (2020).

Family processes are recursive. Conflict between adults can alter caregiving availability, while a child's behavioral difficulties may increase parental strain and disagreement, creating time-varying associations in both directions. Parental mental health, economic pressures, and work circumstances can simultaneously influence conflict and child adjustment. Observational analyses should test reciprocal paths and shared causes rather than presuppose one-way transmission. The genetically sensitive study reported by Harold and colleagues shows why attention to parental psychopathology and inherited liabilities improves interpretation; it also illustrates that genetic control within a specialized sample cannot replace replication across varied families Harold et al. (2012) Erel & Burman (1995).

A child's interpretation is itself developmentally structured. Younger children often depend more heavily on immediate perceptual cues and accessible explanations, whereas older children may integrate family history, anticipate consequences, or feel pressure to mediate between adults. Adolescents can recognize relationship complexity while remaining exposed to loyalty conflicts, economic uncertainty, or disrupted family communication. Developmental stage can modify how exposure is understood and how distress is expressed, but age alone does not indicate vulnerability or resilience. Research needs equivalent constructs and age-appropriate measures before claiming that a specific developmental period is universally most sensitive Grych & Fincham (1990) Grych et al. (1992).

The synthesis uses a multiple-pathway perspective: conflict characteristics shape children's perceptions and regulatory responses; adult stress and behavior may modify parenting and family organization; changes in sleep, emotion, and social participation may affect later adaptation. These are separable hypotheses with overlapping evidence, and a diagram connecting them cannot itself certify each causal arrow. An adequate investigation measures alternative explanations, establishes timing, and identifies the domain in which each mechanism is plausible. The most informative model is one that generates distinct testable predictions—for example, whether child-centered content disproportionately predicts self-blame while adult hostility predicts vigilance and felt insecurity Grych & Fincham (1993) El-Sheikh et al. (2007).

Cognitive Appraisals of Threat, Causality, and Control

The cognitive-contextual framework proposes that children evaluate conflict through the meanings they assign to events, including threat, causal responsibility, anticipated consequences, and perceived ability to cope. These evaluations are situated in the child's knowledge of the relationship, the immediate features of the incident, and the broader social environment. Threat may involve fear of injury, separation, lost caregiving, or persistent family instability; self-blame may involve the belief that one's behavior created the disagreement. These are distinguishable processes because a child can experience high threat without personal responsibility or intense guilt without expecting imminent danger. The 1990 framework specifies the conceptual structure, and subsequent child-report measures allow competing paths to be tested empirically Grych & Fincham (1990) Grych et al. (1992).

Appraisal research places the child's phenomenological viewpoint within a rigorous measurement program. Self-reported perception is evidence about perceived meaning, whereas observation and adult reports describe other aspects of the environment. A child's description of severe hostility may accurately convey subjective distress even when adults describe the event as routine; a parent's account may add information about events the child did not witness. Apparent discrepancies can be analytically valuable. Researchers should avoid treating the informant closest to their preferred theory as automatically correct. Multi-informant designs can estimate both shared exposure and perspective-specific interpretation, and prospective studies can test whether appraisal predicts later symptoms beyond baseline adjustment Grych et al. (1992) Harold & Sellers (2018).

Direct experimental evidence is necessarily limited to ethical manipulations of mild scenarios, vignettes, or simulated interactions. Grych and Fincham's 1993 studies varied conflict content and intensity to examine reported emotion and coping, providing leverage on immediate interpretation. Such tasks illuminate whether children distinguish threatening from less threatening representations; they cannot reproduce repeated exposure to intimidation or violence. Longitudinal correlations between threat appraisal and symptoms offer temporal information but remain susceptible to stable child characteristics, reporting bias, and unmeasured environmental influences. Convergence among experimental, naturalistic, and repeated-measures findings is therefore more informative than treating any single method as definitive Grych & Fincham (1993) Grych & Fincham (1990).

The explanatory value of appraisal is clearest when models specify how change could occur. If credible resolution reduces perceived threat while observable hostility stays constant, a theory predicts altered subsequent emotional responding. If parental reassurance addresses a child's mistaken belief in responsibility, a self-blame pathway predicts improvement in guilt-related reactions without requiring all distress to disappear. Each prediction requires validated measures at appropriate intervals. It is also possible that symptoms change first and alter later appraisals, particularly when anxiety or depression increases threat sensitivity. Research that measures appraisal and adjustment on only one occasion cannot distinguish these directions Grych et al. (1992) van Eldik et al. (2020).

Emotional Security and the Organization of Defensive Responding

Emotional security theory treats the child's sense of safety in the interparental relationship as a developing regulatory concern. Davies and Cummings argued that repeated conflict can influence immediate emotional arousal, behavioral attempts to manage the situation, and internal representations about the relationship's stability. From this perspective, a child's distress during one disagreement is part of a broader history of expectations about whether adult relationships remain safe and predictable. The theory therefore explains why the same overt behavior may carry distinct significance in different families and why reactivity can persist after a particular argument ends. The underlying research question is whether changes in security account for later adjustment above competing family-level pathways Davies & Cummings (1994) Harold & Sellers (2018).

Security and attachment overlap in their concern with safety, but the target relationship differs. Attachment measures commonly assess the child's organization of behavior in relation to a caregiver as a source of support; interparental security concerns the child's experience of the relationship between caregivers. A child may have supportive one-to-one relationships with both parents while remaining apprehensive about their interactions with each other. Conversely, interpersonal conflict may coexist with other strong caregiving resources. Conflating these constructs would obscure the distinct pathways by which adult interaction and caregiving responsiveness are associated with symptoms. The metanalytic dimensions of conflict and children's responses reinforce the need to distinguish external relational conditions from internal security responses Davies & Cummings (1994) van Eldik et al. (2020).

Defensive responses can include vigilance, withdrawal, attempts to intervene, monitoring, heightened reactivity, or disengagement. Their immediate function may be to obtain information, prevent escalation, or reduce exposure. Longitudinal risk emerges when strategies that are understandable during one threatening episode become pervasive, costly, or difficult to regulate across contexts. The same observable withdrawal may reflect adaptive self-protection in an unsafe home and social disengagement that compromises relationships at school. A complete study needs event-level behavior, information about actual danger, and functional outcomes rather than labeling a strategy uniformly maladaptive. Emotional security theory yields testable predictions about the contexts in which these strategies generalize Davies & Cummings (1994) El-Sheikh et al. (2007).

Security processes also require distinction between mediation and mere covariation. If parents' conflict and children's insecurity are recorded simultaneously, and insecurity correlates with symptoms, an indirect statistical association may fit the theory while leaving alternative causal sequences unresolved. Stronger tests employ temporal separation, controls for prior adjustment, independent informants, and analyses that permit bidirectionality. Intervention designs can help when the intervention specifically changes conflict practices and security rather than simultaneously changing many family characteristics. In this way emotional security remains a productive explanatory program whose mechanisms require independently discriminating evidence rather than inference from cross-sectional path diagrams Harold & Sellers (2018) van Eldik et al. (2020).

Triangulation, Loyalty Conflict, and Role Diffusion

Triangulation occurs when a child becomes drawn into an unresolved relationship between adults, for example through recruitment into alliances, messenger roles, secrecy, or pressure to endorse one parent against the other. The child's experience of feeling caught between parents is a distinct construct that can be measured regardless of formal family structure. It differs from ordinary age-appropriate participation in family decisions. The critical elements concern divided loyalties, responsibility for adult emotion, and constrained freedom to sustain relationships. The problem is particularly salient when adults depend on a child to regulate their own interpersonal conflict or transmit hostile communications. Empirical studies of postdivorce families identify role diffusion as a potentially important mechanism linking conflict to child difficulties van Dijk et al. (2020) Schrodt (2026a).

Recent meta-analyses permit greater precision about the magnitude and limits of associations. Schrodt's analysis published in 2026 combined 49 studies with 23,336 participants and found a pooled correlation of .402 between interparental conflict and children feeling caught, with a reported 95 percent confidence interval of .366 to .437. The estimate summarizes a substantial association across included studies; it does not establish that conflict caused triangulation in every family. Variations in measurement, family structure, and informants may affect the average. Researchers should assess whether the child's sense of loyalty conflict is specifically elevated after child-related disputes, direct recruitment, or persistent postseparation hostility Schrodt (2026a).

A related 2026 synthesis examined 54 studies comprising 25,582 individuals. It reported associations between feeling caught and internalizing distress of r = .241, based on 47 studies, and externalizing problems of r = .193, based on 22 studies. These estimates are important because they relate the proposed mechanism to distinct outcome domains rather than simply documenting adult conflict. The pooled correlations are not individual risk probabilities and cannot determine the clinical status of a particular child. Shared reporters and observational designs remain relevant limitations. The resulting empirical picture is consistent with triangulation as a candidate pathway and warrants carefully designed prospective and intervention tests Schrodt (2026b).

Family systems language can illuminate how adult-child roles become reorganized under strain, but it needs behavioral specification. Calling a family enmeshed or dysfunctional supplies little information unless researchers identify observable recruitment practices, constraints on autonomy, communication patterns, and changes in distress. The protection of children's relational freedom is a concrete outcome worthy of measurement. Postseparation research also highlights how role diffusion and hostile parenting can operate together, making it essential to separate the direct impact of feeling caught from the effects of reductions in emotional support. Interventions should address adult communication and decision making without assigning the child responsibility for restoring the couple or coparenting relationship van Dijk et al. (2020) Harold & Sellers (2018).

Spillover into Parenting and Parent–Child Relationships

The spillover hypothesis proposes that emotion, stress, and interaction patterns in the interparental relationship influence the way adults interact with children. Conflict may diminish patience, predictability, or warmth, or increase hostility, psychological intrusiveness, and inconsistent structure. These possibilities differ from the child's direct appraisal of adult conflict, even though both pathways can occur in the same household. Erel and Burman's 1995 meta-analysis of 68 studies reported a moderate positive association between marital and parent-child relationship quality, with an effect size reported as d = .46. The analysis supports the existence of an association while leaving room for bidirectional dynamics, shared causes, and heterogeneous family histories Erel & Burman (1995).

Parenting operates through multiple dimensions. Parental warmth or support concerns availability and responsiveness; hostility concerns harsh, rejecting, or aggressive responses; structure concerns stable expectations and guidance; intrusiveness concerns excessive interference with the child's autonomy; role diffusion concerns inappropriate reassignment of adult functions to the child. A single global parenting index can conceal different or competing processes. The postdivorce synthesis of van Dijk and colleagues examined 115 samples of 24,854 divorced families and modeled parenting dimensions as candidate mediators. Most observed correlations were small and in the expected direction, and several parenting dimensions were implicated in indirect pathways van Dijk et al. (2020).

These meta-analytic mediation results are substantively informative but should be interpreted as a synthesis of associations. Structural equation models fitted to aggregate correlations cannot by themselves prove that changes in conflict precede changes in parenting which then cause changes in adjustment. Some studies may contain cross-sectional designs, shared reporters, or unmeasured characteristics. The distinction matters because the same pattern might arise if a child's difficult behavior increases parental hostility or if an external stressor affects both relationships. Longitudinal models with repeated measures from parents, children, and independent observers can offer stronger tests of direction and developmental change van Dijk et al. (2020) Harold et al. (2012).

An important implication of spillover research is that strengthening caregiving may provide support even when parents' partnership remains strained, provided the environment is safe. The theory does not require treating every family as a couple intervention case or assuming that reunification is a desirable goal. For separated families, respectful coordination, reliable routines, freedom from coercion, and emotionally responsive parent-child ties are distinct potentially modifiable targets. Randomized family and coparenting programs offer evidence about changes in supported domains, while exposure to intimate partner violence calls for different safety-centered strategies. The scientific task is to identify which program component changes which process under which conditions Solmeyer et al. (2014) Jouriles et al. (2009).

Emotional, Behavioral, Social, and Academic Outcomes

Internalizing and externalizing symptoms are the best-studied child-adjustment outcomes in this literature, yet neither is a direct proxy for exposure or a sufficient description of a child's developmental functioning. Internalizing measures commonly include anxious affect, sadness, withdrawal, or somatic concerns; externalizing measures capture aggression, oppositional behavior, and dysregulation. The distinction is useful for hypothesis testing because threat sensitivity, security concerns, and behavioral modeling may have different associations with each domain. However, these problems frequently coexist, and an elevation in a questionnaire score is not equivalent to meeting diagnostic criteria. A 2020 meta-analysis separately estimated associations of relationship dimensions with both classes of difficulty van Eldik et al. (2020) Harold & Sellers (2018).

Children's social relationships and schooling add further dimensions. Worry about conflict may occupy attention during learning, create sleep-related fatigue, or influence peer interactions through irritability and defensive expectations. Families may also experience changes in transportation, housing, routines, or school continuity, particularly after separation, and these practical changes need direct measurement before attributing academic variation to emotional exposure. Studies that examine adjustment solely through parent reports risk combining adult distress with perceptions of the child's behavior. A more defensible design includes teacher observations, child reports, developmental competence indicators, and repeated assessments of educational contexts El-Sheikh et al. (2007) Amato (2001).

Longitudinal adjustment is a trajectory rather than a binary label. Some children exhibit transient symptoms during an acute family transition and recover as household stability improves, whereas others show chronic difficulties or delayed responses. The appropriate analytic task is to estimate how baseline functioning, exposure changes, and subsequent outcomes are related in time. Repeated sampling can distinguish stable between-family differences from within-child changes. Any interpretation of vulnerability should acknowledge that exposure distributions and available support vary; average associations cannot assign a predetermined developmental fate to an individual. Meta-analytic estimates describe population-level relationships that may be moderated by risk, resources, and selection Harold & Sellers (2018) van Eldik et al. (2020).

The research program should also study positive competence rather than only symptoms. Emotional understanding, help-seeking, prosocial behavior, educational engagement, and the ability to maintain safe relationships may be meaningful outcomes. Measuring them separately avoids the assumption that an absence of reported problems means flourishing. The selection of age-appropriate outcomes must be theory-driven: sleep and separation distress may be salient early in development, whereas peer relationships and autonomous conflict management may become increasingly measurable later. Research should compare domains with transparent operational definitions and preserve uncertainty when an effect appears for one adjustment indicator but not another Grych & Fincham (1990) El-Sheikh et al. (2006).

Stress Biology, Sleep, and Regulatory Processes

Interparental conflict may influence child adjustment partly through repeated emotional arousal and its consequences for physiological regulation. The hypothesized pathways include changes in vigilance, autonomic responses, endocrine activity, and restorative sleep, with effects depending on developmental stage and environmental safety. The biological framing should remain specific: a measured cortisol value does not by itself reveal an enduring stress mechanism, and physiological reactivity can reflect adaptive responses to immediate demands. Research is strongest when it integrates repeated biological sampling with independently measured conflict and psychological outcomes. Emotional security theory offers one account of why recurring interpersonal uncertainty could sustain regulatory demands Davies & Cummings (1994) El-Sheikh et al. (2007).

Sleep illustrates a measurable intermediate domain. El-Sheikh and colleagues' 2006 study observed 8- to 9-year-old children using wrist actigraphy across seven consecutive nights, alongside reports of marital conflict and sleepiness. The authors described associations between greater conflict and reduced sleep quantity or quality after adjustment for specified demographic factors. Actigraphy adds an objective movement-based indicator to subjective reports, but observational control of measured variables does not eliminate unobserved confounding. Different sleep parameters can also reflect medical, environmental, and scheduling influences. The findings motivate prospective tests rather than a claim that all exposed children will develop clinically significant sleep disturbance El-Sheikh et al. (2006).

A separate study by El-Sheikh, Buckhalt, Cummings, and Keller examined 166 elementary-school children and tested an indirect model linking marital conflict, children's insecurity, sleep disruptions, and adjustment. The study combined child reports, actigraphy, and assessments of behavioral, emotional, and academic functioning by parents and teachers. Its results supported the proposed pattern in that sample, while the causal ordering of psychological variables requires attention to design and timing. This multimethod approach is valuable because it reduces reliance on a single respondent and connects perceived family conditions to biological regulation. It also demonstrates how a broad adjustment theory can generate narrower empirical predictions El-Sheikh et al. (2007).

Stress processes should be interpreted in the context of safety, socioeconomic resources, housing stability, and everyday routines. Sleep disruption might precede increased irritability that changes family interactions, making reciprocal modeling important. Physiological indicators require clear protocols about time of sampling, medication, illness, circadian timing, and equipment accuracy. Meta-analyses combining heterogeneous measures may obscure these requirements. An ethically appropriate research agenda would test whether reductions in destructive exposure improve sleep and emotional regulation and whether these changes mediate later benefits, while avoiding invasive measurements that add burden without a clear scientific question El-Sheikh et al. (2006) Harold & Sellers (2018).

Child Development from Early Childhood through Adolescence

Early childhood is characterized by emerging emotion regulation and substantial dependence on caregivers for environmental predictability and reassurance. Young children may register loudness, facial expressions, abrupt departures, and changes in caregiver availability before they can explain the motives underlying conflict. Experimental observations of children responding to angry adult behavior helped establish that young children react to differences in adult emotional signals and prior family hostility. These findings do not imply that a single exposure determines later psychopathology. Instead, they clarify the importance of measuring observable reactions and family histories at developmentally appropriate levels Cummings et al. (1989) Davies & Cummings (1994).

During middle childhood, developing perspective-taking and causal reasoning alter how conflict is evaluated. Children may distinguish differences between adults' views yet still conclude that they caused disputes involving schoolwork, behavior, or care arrangements. The CPIC was initially studied among nine- to twelve-year-olds, making its early evidence particularly relevant to this period. Instruments validated in one age range cannot automatically be used with preschool children or adolescents without testing comprehension and measurement structure. Researchers should examine whether conflict-property, threat, and self-blame dimensions retain their meanings across age groups and whether the relevant outcomes change with developmental tasks Grych et al. (1992) Grych & Fincham (1993).

Adolescence introduces increased autonomy, more complex family knowledge, and expanded social relationships. Exposure can involve direct observation, digital messages, parental disclosures, or obligations to relay communications between households. A young person may offer practical assistance while simultaneously experiencing pressure to make adult decisions. Triangulation becomes particularly important when relationship autonomy is constrained by conflicting parental demands. The 2026 quantitative syntheses of feeling caught bring evidence from multiple developmental groups into a shared research program, although pooled estimates cannot establish identical processes at every age. Age-related moderation must be tested within studies with adequate coverage Schrodt (2026a) Schrodt (2026b).

Developmental continuity and change are both plausible. An early pattern of heightened vigilance can persist, diminish with improvements in family safety, or reappear around new transitions. Retrospective adult recollections may illuminate perceived histories but are vulnerable to reconstruction and selective recall. Prospective cohorts following children across family transitions can provide more robust evidence about timing, cumulative exposure, and the circumstances that support adaptation. Cross-lagged associations should be interpreted with attention to the separation of stable family differences from within-person processes. A credible developmental account must explain changing risks and opportunities without assuming a universal age-specific causal sequence Harold & Sellers (2018) Amato (2001).

Family Separation, Divorce, and Postseparation Conflict

Divorce is a family transition with possible changes in household composition, resources, routines, parental contact, and legal arrangements. Interparental conflict is a relationship process that may occur before, during, or after separation, and it can be present in continuously partnered families. These constructs must be distinguished in both theory and measurement. Studies comparing children of divorced and continuously married parents may identify differences across populations while leaving the contributions of preceding conflict, selection, economic change, and continuing parenting practices unresolved. Amato's 2001 meta-analytic update documented average differences across several adjustment indicators and exemplified both the value and limitations of structural comparisons Amato (2001) Harold & Sellers (2018).

The 2020 synthesis by van Dijk, van der Valk, Deković, and Branje focused specifically on divorced families. It compiled 2,257 correlations across 115 samples representing 24,854 families and used three-level meta-analysis together with meta-analytic structural equation models. The authors reported mostly small associations among interparental conflict, parenting dimensions, and child psychosocial adjustment. Several parenting features appeared as mediating mechanisms within the statistical models, with role diffusion highlighted as a particular concern. The analysis offers a useful map of relationships among constructs, but a statistical indirect path assembled from observational studies should not be equated with a verified causal chain van Dijk et al. (2020).

Persistent postseparation hostility may create uncertainty about schedules, emotional availability, and permission to maintain relationships with both caregivers. Some children experience arguments during handoffs, hostile messages, or requests to deliver information. Others have little direct contact with parental disagreement because adults maintain stable boundaries. These differences show why a label such as divorced family is too coarse for mechanism research. Studies should record the quality and safety of contact, the child's awareness of conflict, and contextual resources, rather than treating custody arrangement or household form as intrinsically explanatory. Severe coercive control requires a safety-specific account that cannot presume routine shared decision making van Dijk et al. (2020) World Health Organization (2026).

A nuanced interpretation of postseparation adjustment combines conflict exposure, disrupted or sustained caregiving, economic transitions, and the child's developmental perspective. Because adults and children may report the same transition differently, independent informants are particularly important. Family interventions can target predictable routines and supportive caregiving without requiring a renewed intimate partnership, and legal or service environments may shape what changes are feasible. Scientific conclusions should be tied to observed processes rather than moral judgments about household structure. The key question is which modifiable exposures predict improvements in safety, security, and functioning when family organization changes Amato (2001) Solmeyer et al. (2014).

Measurement of Interparental Conflict and Child Adjustment

Measurement begins with an explicit definition of the construct and the observational unit. Frequency counts describe how often a specified event occurs; ratings of intensity describe its perceived seriousness; behavioral coding characterizes hostility, withdrawal, repair, and child involvement. These indicators should not be reduced to one score without empirical justification because different conflict dimensions can show distinct relationships with outcomes. The Children's Perception of Interparental Conflict Scale was developed to measure the child's view of conflict properties, threat, and self-blame. In its early validation study, factor-analytic subscales demonstrated acceptable reliability, but the initial samples and measurement structure must be considered when extending the instrument to other ages and populations Grych et al. (1992) van Eldik et al. (2020).

Parent reports and child reports answer related but different questions. Adult ratings may better capture private disagreements, recurrent relationship dissatisfaction, or the intent behind interaction; child reports capture witnessed events and personal interpretations. Independent observers can code observable behavior without access to internal experience. When the same child reports both conflict threat and anxiety, the resulting association partly reflects the common method of measurement, while differences between reporters may reflect real context specificity. Studies should explicitly model reporter effects, use age-appropriate scales, and conduct longitudinal measurement-invariance tests before interpreting changes in scores as changes in the underlying construct Grych et al. (1992) Harold & Sellers (2018).

Adjustment measurement also demands domain specificity. A validated measure of internalizing symptoms should not be treated as equivalent to a clinical diagnosis; externalizing indicators require distinctions among impulsivity, aggression, and norm-violating behavior. School records, teacher reports, sleep actigraphy, clinician assessments, and direct observation may provide complementary evidence. The value of independent measures is illustrated by studies relating reported parental conflict to actigraphy-based sleep disruption and informant-rated functioning. Every measure has its own validity boundary: sleep movement can be measured objectively without revealing the meaning of a child's distress or providing a direct index of interpersonal safety El-Sheikh et al. (2006) El-Sheikh et al. (2007).

A rigorous instrument audit reports internal consistency, test-retest stability, dimensional structure, discrimination among constructs, age and cultural validation, and sensitivity to change. Researchers should preserve separate scores if theory predicts different pathways from child-related conflict, hostility, and disengagement. If an intervention claims improvement, the outcome instrument should detect change in a target process rather than merely reflect general parental optimism. Scale translations require evidence of linguistic and psychometric equivalence. Measurement development is thus part of causal theorizing: vague exposure categories weaken every later claim about mediation, moderation, treatment response, and cumulative risk van Eldik et al. (2020) Grych et al. (1992).

Experimental, Observational, and Intensive Longitudinal Designs

Laboratory experiments can vary the content or form of ethically mild interpersonal disagreements and observe children's immediate affective, cognitive, or behavioral reactions. The experimental advantage is control over specified features of the stimulus; the principal limitation is that constructed scenes may differ from the child's lived history, and genuine threatening exposure cannot ethically be assigned. Early observational and experimental studies of children's responses to adult anger demonstrate the usefulness of standardized episodes and reconciliation cues. Their results must be interpreted as responses within a particular experimental context, not as direct proof of long-term effects in every family Cummings et al. (1989) Grych & Fincham (1993).

Naturalistic observation captures spontaneous interaction, including interruptions, repair attempts, nonverbal cues, and how children enter or leave a conflict. Video-based coding requires clear behavior categories, coders trained to criterion, and assessment of interrater agreement. Families who consent to being filmed may differ from families experiencing coercive danger, limiting generalization. Observed sequences can establish temporal order within an episode, such as whether an adult expression preceded a child's approach or withdrawal, but they may not establish why the interaction occurred. Combining behavioral observation with children's private perceptions and later outcomes supports a more complete account of the processes Harold & Sellers (2018) van Eldik et al. (2020).

Longitudinal studies offer a developmental view by assessing conflict, parenting, appraisals, and adjustment at multiple occasions. They can estimate whether prior conflict predicts subsequent change while adjusting for baseline symptoms and measured contextual variables. Their conclusions depend on the number of waves, spacing between waves, attrition, stability of measures, and modeling assumptions. Conventional cross-lagged models can conflate enduring differences among families with changes inside one family. Designs that explicitly separate these sources of variance, and examine bidirectional associations, are preferable when the research question concerns a within-family mechanism. Even sophisticated observational models remain vulnerable to unmeasured changes occurring between assessments Harold et al. (2012) van Dijk et al. (2020).

Intensive longitudinal methods, including event-based diaries and repeated brief assessments, can measure the sequence of conflict, immediate child appraisal, sleep, and next-day functioning. They also raise practical and ethical challenges: repeated questioning can burden participants, a child's privacy may be compromised, and recording household conflict may create danger in coercive settings. Study protocols should define safe contact and data confidentiality while selecting a sampling density justified by the proposed mechanism. Within-person analyses can test whether deviation from a family's usual conflict level predicts changes in reported security or behavior, providing a stronger match between theoretical timing and empirical observation El-Sheikh et al. (2007) Harold & Sellers (2018).

Meta-Analytic Evidence and Interpretation of Effect Sizes

A meta-analysis estimates associations across studies that differ in samples, measurement, and design. van Eldik and colleagues' 2020 Psychological Bulletin analysis incorporated 169 studies relevant to child maladjustment and 61 studies relevant to children's responses to conflict. By differentiating relationship quality, conflict frequency, hostile conflict, disengagement, unconstructive conflict, and child-related conflict, it provided a more discriminating synthesis than analyses treating exposure as a single variable. The abstract reports support for some theoretical expectations and challenges to others. Interpretations should preserve this complexity rather than translate the review into the blanket proposition that every disagreement damages children van Eldik et al. (2020).

The postdivorce meta-analysis by van Dijk and colleagues offers a complementary analytical focus. Its 115 samples and 24,854 families were selected for a more specific population and research question. Three-level meta-analytic models accounted for multiple correlations within studies, while meta-analytic structural equation models examined parenting as a possible indirect pathway. The mostly small associations suggest that cumulative family processes can matter without implying that a single exposure yields a large predictable individual outcome. Moreover, the evidence pertains to studies of divorced families, and transferring estimated pathways unchanged to every partnered household requires additional research van Dijk et al. (2020).

Effect sizes must be compared only when they represent compatible estimands. Erel and Burman's reported d = .46 referred to a marital–parent-child relationship linkage, not a direct intervention effect on childhood disorder. Schrodt's 2026 correlation of .402 quantified conflict and feeling caught, while the separate correlations of .241 and .193 linked feeling caught to internalizing and externalizing outcomes. These numerical summaries concern different constructs and should not be ranked as though they were competing measures of one mechanism. Confidence intervals describe sampling uncertainty under the underlying models; they do not automatically capture every form of measurement bias or confounding Erel & Burman (1995) Schrodt (2026a) Schrodt (2026b).

The interpretation of pooled results should account for overlapping primary studies, publication bias, heterogeneity, age distributions, geographic coverage, and the reliance on observational correlations. In particular, meta-analytic mediation may be consistent with a pathway without verifying its causal direction. Reviews should specify which reports were available in full text and which findings were checked from abstracts or bibliographic records. This monograph does not claim a new systematic search, PRISMA-compliant review, or pooled effect estimate. Its numerical examples are quoted from identifiable source reports to illustrate the evidential distinctions needed for a scientifically defensible synthesis van Eldik et al. (2020) van Dijk et al. (2020).

Genetic, Environmental, and Reciprocal Confounding

A central inferential difficulty is that parental relationship quality, parental psychopathology, parenting practices, and child symptoms can share underlying determinants. Inherited dispositions may affect both adult and child characteristics, while economic insecurity, illness, social isolation, or exposure to violence may influence conflict and adjustment simultaneously. Statistical adjustment for a few demographic variables does not necessarily remove these dependencies. The research question is therefore causal and relational: whether changing interparental conflict itself would change child outcomes, and how much of any effect operates through parenting, interpretation, or biological regulation. Answering it demands research designs beyond a single observational correlation Harold et al. (2012) van Eldik et al. (2020).

Harold and colleagues' 2012 study used a genetically sensitive in vitro fertilization design that compared genetically related and unrelated parent-child groupings. It examined interparental conflict, parental psychopathology, hostile parenting, and child antisocial behavior while considering the potential influence of shared genetic factors. This design provides leverage on a confounding problem that conventional biological-family samples cannot isolate as readily. Its contribution is not proof of one universal effect, since assisted reproduction samples may differ on demographic and clinical characteristics. The strongest conclusion is that genetically informed analysis can refine the interpretation of environmental pathways and should be considered alongside prospective and experimental evidence Harold et al. (2012).

Reciprocal influence creates a further challenge. A child experiencing behavioral difficulties may increase caregiving demands and parental disagreement; a worsening parental relationship can then change the child's environment. These feedback patterns can coexist with direct conflict-related stress. Longitudinal models should measure both directions and report how estimates change under alternative lag structures, stable-trait controls, and missing-data assumptions. Natural experiments involving external changes in family conditions may contribute evidence if their assumptions are defensible. The key is not to require that the relationship have a single cause but to identify which processes remain distinguishable when competing explanations are evaluated Harold & Sellers (2018) van Dijk et al. (2020).

Selection and reporting mechanisms also shape published findings. Families who volunteer for relationship interventions may be more motivated or less endangered than families excluded for safety reasons. Children whose teachers notice externalizing problems may be referred disproportionately to services, while quieter internalizing difficulties remain underidentified. Attrition from longitudinal samples can correlate with both conflict severity and psychosocial adversity. Credible studies should report sample recruitment, exposure distribution, exclusion rules, missingness, sensitivity analyses, and whether the same explanatory model survives alternative measurement choices. Causal interpretation is a conclusion earned through converging evidence, not an attribute conferred by elaborate statistical terminology Kitzmann et al. (2003) Harold et al. (2012).

Prevention and Coparenting Intervention Evidence

Intervention research provides an opportunity to examine whether changing family interaction processes improves children's functioning. Preventive approaches often focus on constructive communication, emotional regulation, coordination between caregivers, and consistent support. A randomized study can establish the effect of assignment to a specified program relative to the tested comparator under its implementation conditions. It cannot automatically isolate which component produced a benefit if several were delivered together. The Family Foundations program is relevant because it was designed around coparenting during the transition to parenthood, a period in which relationship patterns and caregiving demands are changing together Solmeyer et al. (2014) Feinberg et al. (2014).

Solmeyer and colleagues' 2014 report tested whether coparenting processes statistically mediated effects of Family Foundations on child adjustment. It offers a more informative question than merely asking whether participating families reported higher satisfaction, because it relates a program target to downstream outcomes. Yet mediated intervention effects need close attention to intervention timing, postrandomization confounding, reliability of the mediator, and the time at which child functioning is assessed. Assignment to treatment may be randomized while the mediator itself is not. The study therefore strengthens the intervention-oriented evidence base without closing every causal question about the precise mechanism Solmeyer et al. (2014).

Feinberg and colleagues published a 2014 long-term follow-up of a randomized Family Foundations trial examining emotional, behavioral, and school adjustment. Longer observation matters because early changes in parental coordination may have delayed implications, and initial gains can dissipate as family circumstances change. A sound interpretation considers the original allocation, attrition, outcome informants, multiplicity of endpoints, and whether the program population represents families experiencing the most severe violence. Findings from preventive coparenting services should not be generalized to situations in which contact or joint planning would jeopardize a survivor's safety. The evidence supports studying specific preventive mechanisms in defined populations Feinberg et al. (2014) World Health Organization (2026).

Cummings and colleagues' 2008 study of a brief program targeting marital conflict in community families offers another prevention example. Its clinical and developmental relevance lies in the possibility of addressing harmful interaction patterns before severe distress becomes entrenched, while keeping participation demands feasible. The outcome of interest must be specified: changed adult conflict practices, observed resolution, children's security, or long-term adjustment are distinct endpoints. Trials need independent outcome assessment and transparent reporting of adverse events as well as average benefits. For future programs, the most valuable comparison would separate instruction in conflict repair from support for parenting and evaluate which components produce durable, developmentally meaningful changes Cummings et al. (2008) Harold & Sellers (2018).

Clinical and Safeguarding Responses to Violence

Children exposed to intimate partner violence face a risk environment fundamentally different from the ordinary disagreement of safe caregivers. Such environments may include physical danger, coercive surveillance, threats, severe intimidation, and economic restrictions. Exposed children can be affected even without being the direct target of an assault, and direct child maltreatment can coexist with partner violence. Research and practice must assess these experiences separately. The World Health Organization identifies witnessing violence between caregivers as a childhood risk factor and emphasizes multisectoral prevention and protection. The scientific literature also recognizes substantial heterogeneity in outcomes and methods World Health Organization (2026) Kitzmann et al. (2003).

Kitzmann and colleagues' 2003 review assembled 118 studies examining psychosocial outcomes among children exposed to domestic violence. The synthesis found associations between exposure and problems relative to comparison groups, while the size of the association depended partly on design and comparison definition. Observational studies in this domain face challenges of overlapping childhood adversities, differential access to services, and the difficulty of recruiting safe, representative samples. It would be unjustified to interpret a pooled effect as a destiny for every child, or to infer the relative safety of two households from relationship status alone. Risk evaluation must focus on actual behavior, danger, and access to support Kitzmann et al. (2003).

Project Support was evaluated in randomized trials involving families affected by intimate partner violence. Jouriles and colleagues' 2009 clinical trial examined reductions in conduct problems among exposed children, offering intervention evidence in a population for whom ordinary couple communication programs may be inappropriate. The conceptual target included children's behavioral functioning and support for the nonabusive caregiver. Randomization improves conclusions about the tested program, while sample criteria, comparator services, practitioner training, and follow-up duration determine the extent of generalization. Research should also monitor whether participation increases risk, and programs should protect children and survivors rather than demand reconciliation or shared conflict responsibility Jouriles et al. (2009) World Health Organization (2026).

Clinical interpretation requires independent assessment of children's symptoms, functional impairment, developmental history, ongoing safety, and available caregiving resources. Screening results can guide evaluation but do not substitute for diagnosis. Interventions may involve trauma-informed care, family support, survivor services, and practical arrangements that reduce exposure. A service plan suitable for one family may be unsafe in another because coercive control changes the meaning of negotiation and contact. The research literature supports tailoring interventions to identifiable exposures and outcomes, with qualified clinical review for detailed practice recommendations. No individual diagnosis, medication regimen, or emergency safety decision follows from a general research synthesis Jouriles et al. (2009) Harold & Sellers (2018).

Resilience, Protective Factors, and Heterogeneous Outcomes

Resilience is studied as the maintenance or recovery of adaptive functioning under adversity, with attention to changing resources and developmental circumstances. It should be operationalized through trajectories and competencies rather than inferred from an impression that a child appears calm. Stable nurturing relationships, access to supportive adults, predictable routines, safe housing, and opportunities for peer engagement may provide resources that interact with conflict exposure. Their benefits must be evaluated rather than presumed: a protective association can reflect selection, prior functioning, or unmeasured social advantage. Reviews of interparental conflict recognize that average risk does not imply a uniform child outcome Harold & Sellers (2018) van Eldik et al. (2020).

Protection can also be domain-specific. A supportive teacher may sustain academic engagement without changing uncertainty about the caregivers' relationship. A trusted grandparent may offer emotional availability while household conflict continues. Reliable care from one parent may strengthen security in that relationship, yet the child can still experience fear regarding interparental violence. Researchers should measure each outcome and resource separately to avoid assuming that improvement in one domain resolves every problem. Theories of family spillover predict that maintaining warmth and structure may reduce some secondary effects, while cognitive-contextual theories predict that clear, age-appropriate information might alter particular appraisals Erel & Burman (1995) Grych & Fincham (1990).

Protective factors may function as statistical moderators, mediators, or outcomes themselves. For example, parental support could moderate the strength of an exposure-symptom association; alternatively, the quality of support may deteriorate in response to conflict, forming part of an indirect pathway. These models have different implications for the timing of measurement and design of intervention. A moderator identified in a small subgroup analysis should be treated cautiously because multiple testing, restricted measurement ranges, and sampling variation can create unstable results. Stronger evidence requires preregistered hypotheses, adequate interaction power, and replication across independent samples van Dijk et al. (2020) Harold & Sellers (2018).

An ecological perspective also demands attention to housing, income, neighborhood safety, discrimination, access to health care, and school resources. These conditions can constrain caregivers' options for separation, treatment, and conflict resolution while shaping children's daily functioning. Interventions aimed exclusively at interpersonal skills may underperform when material conditions remain unstable. At the same time, broad social context should not obscure the particular behavior of caregivers, including violence and efforts at repair. A cumulative theory must be able to identify multiple interacting sources of support and risk without reducing them to a single family quality score Harold & Sellers (2018) World Health Organization (2026).

Cultural Context, Family Diversity, and Measurement Transportability

Research populations differ in family structures, languages, residential arrangements, migration histories, and social expectations about parental roles. The label interparental encompasses biological parents, adoptive parents, stepparents, separated caregivers, and other adults who share consequential responsibility for a child. Studies restricted to married heterosexual couples provide evidence about those sampled relationships; their results must be examined before being applied to other family organizations. Constructive conflict, autonomy, family obligation, and acceptable disclosure may be interpreted differently across settings, making behavioral descriptions and child perspectives preferable to unexamined cultural assumptions Harold & Sellers (2018) van Eldik et al. (2020).

Cultural generalization requires more than translating questionnaire wording. A scale measuring threat or self-blame must be understandable across language groups and demonstrate comparable factor structure, thresholds, and score interpretation. Adults and children may use different norms when describing intensity, warmth, or resolution, and survey items may fail to represent culturally specific forms of relational support or withdrawal. Differences in observed averages can be artifacts of measurement, recruitment, or socioeconomic conditions unless these possibilities are investigated. The original CPIC validation among American school-aged children is an important historical reference, not universal proof of psychometric equivalence Grych et al. (1992).

Family reorganization also intersects with institutions. Legal systems govern custody and protection; schools and health services vary in availability; migration and language barriers can complicate access to care. These conditions affect which children enter research samples and which intervention participants can remain in follow-up. Cross-national comparisons should describe structural conditions instead of treating country as an explanatory mechanism in its own right. Partnerships between families, community organizations, and researchers can improve ecological relevance, while independent scientific reporting remains necessary to prevent local practices from being represented as universally effective without supporting outcome studies World Health Organization (2026) Harold & Sellers (2018).

A scientifically rigorous multicultural program would recruit diverse families prospectively, test measurement invariance, report forms of exposure and safety separately, and examine whether established pathways reproduce across contexts. It would also investigate counterexamples: children with high exposure who maintain functioning and children who experience distress despite low observed overt conflict. Such heterogeneity can identify unmeasured contextual features or distinct developmental routes. The goal is transportable explanations defined through observable processes, not the extension of historically dominant family norms to all populations. Generalization is an empirical claim that should be tested through sampling, replication, and mechanistic comparison van Eldik et al. (2020) Schrodt (2026b).

Comparing Competing Explanatory Pathways

The principal theories make overlapping but discriminable claims. Cognitive-contextual models predict effects through children's appraisal of threat, causation, and coping resources; emotional security models emphasize developing expectations about the safety of interparental relations and the organization of defensive responding. Spillover models predict that adult conflict changes parenting behavior, whereas family systems accounts examine triangulation, altered boundaries, and role allocation. Stress-regulation models address physiological and sleep processes. These mechanisms can coexist without becoming interchangeable. Scientific integration requires distinct measurements, temporal hypotheses, and tests that compare explanatory performance rather than treating agreement with a general family-risk narrative as sufficient evidence Grych & Fincham (1990) Davies & Cummings (1994) Erel & Burman (1995).

One discriminating scenario concerns a child who witnesses a conflict resolved constructively while parenting remains consistently supportive. The appraisal model would examine how the resolution changes perceived threat; the security account would ask whether repeated resolution fosters expectations of relational stability; spillover should predict limited change in parenting if parent-child behavior is unaffected. Another scenario involves hostile adult exchanges outside the child's presence that reduce caregiver availability. In that setting, a parenting pathway may operate even if the child reports little direct exposure. These examples are theoretical tests, not claims that particular interventions have experimentally isolated every component van Eldik et al. (2020) van Dijk et al. (2020).

A third scenario involves postseparation adults whose contact is minimal but whose messages repeatedly recruit the child as mediator. The triangulation framework predicts a distinctive burden associated with divided loyalties, and available meta-analytic evidence supports an association between feeling caught and both internalizing and externalizing difficulties. The existence of a moderate conflict–feeling-caught correlation does not establish that the latter completely explains the former. Comparative models should include direct exposure, family safety, parenting, and reciprocal child effects. They should report the consequences of alternative measurement and confounding assumptions rather than announce a preferred mechanism from one statistical pathway Schrodt (2026a) Schrodt (2026b).

This comparative synthesis favors a multilevel research architecture rather than a single global coefficient. Exposure is characterized by observable conflict dimensions; mechanisms are defined through psychological appraisal, security, parental behavior, role processes, and biological regulation; outcomes are measured across developmentally meaningful domains. Contextual and child-level moderators condition the strength and form of pathways. An empirically useful account specifies when each path should weaken, strengthen, or disappear, and what observations would count against it. This organization preserves the cumulative achievements of theory and meta-analysis while making the next generation of causal research more discriminating Harold & Sellers (2018) van Eldik et al. (2020).

Research Ethics, Clinical Interpretation, and Data Stewardship

Research involving children and conflict-sensitive family information requires specific safeguards for consent, assent, confidentiality, and emotional safety. A child might disclose ongoing threats unknown to the interviewing caregiver, and parents may seek access to responses that could expose the child to retaliation. Data-collection procedures should therefore address private participation, mandated reporting obligations within the relevant jurisdiction, secure communications, and safe withdrawal. These issues are particularly acute in studies of intimate partner violence, where recording a conversation or sending reminders can create exposure. Sound scientific design incorporates safety rather than treating it as an external administrative formality World Health Organization (2026) Jouriles et al. (2009).

Clinical interpretation should separate a developmental risk indicator from a diagnosis. A report of elevated emotional insecurity or a high interparental conflict score indicates a measured perception or exposure, not a particular mental disorder. Diagnostic evaluation requires assessment of symptoms, impairment, duration, alternative explanations, and developmental context. Researchers should distinguish the prevalence of exposure from the prevalence of disorders among exposed children. General literature on child adjustment cannot justify individualized treatment decisions or assumptions that a specific caregiver is responsible for a child's symptoms. When findings concern care, their evidential strength must be matched to the sample and intervention actually studied Harold & Sellers (2018) Kitzmann et al. (2003).

Ethical reporting also concerns terminology. Describing violence as mutual conflict can erase coercion and confuse responsibility, while treating every parental disagreement as traumatic can exaggerate population findings. Researchers should report observable behavior, power asymmetry, safety conditions, and the child's knowledge of events. Family diversity must be represented without assigning risk to a household form in the absence of evidence about actual processes. Publications should disclose recruitment sources, exclusions, conflicting results, and the limitations of extrapolating from intervention volunteers to families facing severe material or safety constraints World Health Organization (2026) Amato (2001).

Digital family communication creates further methodological questions: caregivers may argue through messaging platforms, schedule exchanges through applications, or ask children to forward emotionally consequential material. In the Artificial Era, automated systems may mediate scheduling and communication or shape how adults seek relationship guidance. These possibilities justify research on exposure, privacy, accountability, and whether technological mediation increases or reduces children's involvement in adult disputes. No finding in the reviewed historical interparental-conflict literature establishes a distinctive causal effect of artificial agents. Such hypotheses require independently specified exposures, prospective data, comparative designs, and protection against unauthorized collection of children's information Harold & Sellers (2018) Schrodt (2026a).

Unresolved Questions and a Cumulative Research Agenda

The strongest remaining question concerns causation across time. Existing meta-analyses consistently organize correlations between interparental relationship quality, parenting, children's responses, and adjustment, but the relative contribution of each pathway depends on design and measurement. Future cohorts should begin before major family transitions, assess baseline child and caregiver functioning, measure specific conflict episodes repeatedly, and use independent outcomes. Genetically sensitive and quasi-experimental approaches can complement randomized preventive interventions. Mechanism claims should specify the estimand and identify the conditions under which indirect effects are causally interpretable van Eldik et al. (2020) van Dijk et al. (2020) Harold et al. (2012).

A second problem is the granularity of exposure. Research should distinguish hostile escalation, child-related disagreement, disengagement, credible repair, and coercive danger; it should also measure whether children see an argument, hear it indirectly, or infer its continuation. The meaning of an episode may depend on previous experiences and family safety. Event-level designs paired with longer developmental follow-up could establish whether immediate appraisals predict cumulative security changes and whether changes in sleep explain any subsequent symptoms. Well-defined measures make it possible to test mechanisms that would otherwise be buried inside an overall marital discord score Grych et al. (1992) El-Sheikh et al. (2007).

A third issue concerns the translation from evidence to prevention. Trials should specify which conflict dimension or parenting process is targeted, compare active ingredients when feasible, and evaluate outcomes that matter to children rather than only parental satisfaction. Intervention research requires longer follow-up, independent assessors, evaluation of unintended harms, and reporting of inclusion and exclusion rules. Safety-sensitive programs must be evaluated separately from couple enrichment services. The evidence from Family Foundations, brief conflict-prevention programs, and Project Support illustrates promising but distinct settings, populations, and mechanisms Feinberg et al. (2014) Cummings et al. (2008) Jouriles et al. (2009).

The final research priority is reproducible heterogeneity analysis. Larger multi-site samples should test measurement equivalence across developmental stages, cultures, and family forms, then distinguish within-family change from between-family differences. Data-sharing practices must respect children's privacy and violence-related safety. Open methods, clearly described analytic decisions, and replication across independent teams will improve the interpretation of moderators and mediated pathways. The scientific ideal is an account that explains why some children suffer persistent difficulties, why others maintain functioning, and which observable changes improve safety and adaptation within each relevant context Harold & Sellers (2018) van Eldik et al. (2020) Schrodt (2026b).

Conclusion

Interparental conflict and child adjustment constitute a developmental psychological research field concerned with the consequences of adult relational processes for children's security, interpretation, regulation, and adaptation. The strongest evidence establishes systematic associations between distinguishable dimensions of interparental relations and children's emotional and behavioral difficulties. Foundational theories provide complementary pathways through threat appraisal, emotional security, parenting spillover, and family role reorganization. Their value lies in the distinct observable predictions they produce, particularly when the child's own interpretation is measured alongside adult behavior and longer-term functioning Grych & Fincham (1990) Davies & Cummings (1994) van Eldik et al. (2020).

Quantitative reviews establish an important cumulative base: broad interparental relationship findings, postdivorce parenting mediation patterns, and specific associations involving children feeling caught between caregivers. Experimental observation clarifies immediate responses to conflict cues, while prospective studies, sleep measurement, and genetically sensitive research strengthen the analysis of potential mechanisms. Randomized programs demonstrate that aspects of family functioning can be targeted for prevention or treatment in defined populations. Yet observational mediation remains distinct from experimental identification, and each numerical result belongs to its particular construct, sample, and design van Dijk et al. (2020) Schrodt (2026a) Schrodt (2026b) Feinberg et al. (2014).

The central substantive conclusion is that the developmental meaning of parental disagreement depends on how adults conduct it, whether resolution and safety are credible, how children understand the event, and what support remains available. Children who encounter destructive conflict may respond through vigilance, self-blame, divided loyalties, altered caregiving relationships, and dysregulated sleep or emotion. These processes are conditional, mutually interacting, and shaped by family histories and external resources. Violence introduces a specific obligation to study danger and protection rather than assume that relationship repair is the appropriate endpoint Kitzmann et al. (2003) World Health Organization (2026) El-Sheikh et al. (2007).

A complete scientific account therefore integrates relationship behavior, child perspective, family organization, biological regulation, and development over time while respecting their distinct methods and evidential boundaries. Future progress depends on precise exposure definitions, multi-informant longitudinal measurement, ethically designed causal tests, independently evaluated interventions, and transparent generalization across family contexts. Research on new technological mediators can extend these questions if it begins with the same operational discipline. The resulting field remains focused on a concrete scientific problem: the conditions under which children can maintain safety, competence, and psychological continuity amid changing relationships between their caregivers Harold & Sellers (2018) van Eldik et al. (2020).

References

  • Amato, P. R. (2001). Children of divorce in the 1990s: An update of the Amato and Keith (1991) meta-analysis. Journal of Family Psychology, 15(3), 355–370. https://doi.org/10.1037/0893-3200.15.3.355
  • Cummings, J. S., Pellegrini, D. S., Notarius, C. I., & Cummings, E. M. (1989). Children's responses to angry adult behavior as a function of marital distress and history of interparent hostility. Child Development, 60(5), 1035–1043. https://doi.org/10.1111/j.1467-8624.1989.tb03534.x
  • Cummings, E. M., Faircloth, W. B., Mitchell, P. M., Cummings, J. S., & Schermerhorn, A. C. (2008). Evaluating a brief prevention program for improving marital conflict in community families. Journal of Family Psychology, 22(2), 193–202. https://doi.org/10.1037/0893-3200.22.2.193
  • Davies, P. T., & Cummings, E. M. (1994). Marital conflict and child adjustment: An emotional security hypothesis. Psychological Bulletin, 116(3), 387–411. https://doi.org/10.1037/0033-2909.116.3.387
  • El-Sheikh, M., Buckhalt, J. A., Mize, J., & Acebo, C. (2006). Marital conflict and disruption of children's sleep. Child Development, 77(1), 31–43. https://doi.org/10.1111/j.1467-8624.2006.00854.x
  • El-Sheikh, M., Buckhalt, J. A., Cummings, E. M., & Keller, P. (2007). Sleep disruptions and emotional insecurity are pathways of risk for children. Journal of Child Psychology and Psychiatry, 48(1), 88–96. https://doi.org/10.1111/j.1469-7610.2006.01604.x
  • Erel, O., & Burman, B. (1995). Interrelatedness of marital relations and parent-child relations: A meta-analytic review. Psychological Bulletin, 118(1), 108–132. https://doi.org/10.1037/0033-2909.118.1.108
  • Feinberg, M. E., Jones, D. E., Roettger, M. E., Solmeyer, A., & Hostetler, M. L. (2014). Long-term follow-up of a randomized trial of Family Foundations: Effects on children's emotional, behavioral, and school adjustment. Journal of Family Psychology, 28(6), 821–831. https://doi.org/10.1037/fam0000037
  • Grych, J. H., & Fincham, F. D. (1990). Marital conflict and children's adjustment: A cognitive-contextual framework. Psychological Bulletin, 108(2), 267–290. https://doi.org/10.1037/0033-2909.108.2.267
  • Grych, J. H., Seid, M., & Fincham, F. D. (1992). Assessing marital conflict from the child's perspective: The Children's Perception of Interparental Conflict Scale. Child Development, 63(3), 558–572. https://doi.org/10.1111/j.1467-8624.1992.tb01646.x
  • Grych, J. H., & Fincham, F. D. (1993). Children's appraisals of marital conflict: Initial investigations of the cognitive-contextual framework. Child Development, 64(1), 215–230. https://doi.org/10.1111/j.1467-8624.1993.tb02905.x
  • Harold, G. T., Elam, K. K., Lewis, G., Rice, F., & Thapar, A. (2012). Interparental conflict, parent psychopathology, hostile parenting, and child antisocial behavior: Examining the role of maternal versus paternal influences using a novel genetically sensitive research design. Development and Psychopathology, 24(4), 1283–1295. https://doi.org/10.1017/S0954579412000703
  • Harold, G. T., & Sellers, R. (2018). Annual research review: Interparental conflict and youth psychopathology: An evidence review and practice focused update. Journal of Child Psychology and Psychiatry, 59(4), 374–402. https://doi.org/10.1111/jcpp.12893
  • Jouriles, E. N., McDonald, R., Rosenfield, D., Stephens, N., Corbitt-Shindler, D., & Miller, P. C. (2009). Reducing conduct problems among children exposed to intimate partner violence: A randomized clinical trial examining effects of Project Support. Journal of Consulting and Clinical Psychology, 77(4), 705–717. https://doi.org/10.1037/a0015994
  • Kitzmann, K. M., Gaylord, N. K., Holt, A. R., & Kenny, E. D. (2003). Child witnesses to domestic violence: A meta-analytic review. Journal of Consulting and Clinical Psychology, 71(2), 339–352. https://doi.org/10.1037/0022-006X.71.2.339
  • Schrodt, P. (2026). Interparental conflict and parent–child triangulation: A meta-analytical review of children feeling caught between parents. Human Communication Research, 52(2), 77–89. https://doi.org/10.1093/hcr/hqaf018
  • Schrodt, P. (2026). Parent–child triangulation and children's adjustment: A meta-analysis of children feeling caught and symptoms of distress. Annals of the International Communication Association, advance online publication, Article wlag024. https://doi.org/10.1093/anncom/wlag024
  • Solmeyer, A. R., Feinberg, M. E., Coffman, D. L., & Jones, D. E. (2014). The effects of the Family Foundations prevention program on coparenting and child adjustment: A mediation analysis. Prevention Science, 15(2), 213–223. https://doi.org/10.1007/s11121-013-0366-x
  • van Dijk, R., van der Valk, I. E., Deković, M., & Branje, S. (2020). A meta-analysis on interparental conflict, parenting, and child adjustment in divorced families: Examining mediation using meta-analytic structural equation models. Clinical Psychology Review, 79, Article 101861. https://doi.org/10.1016/j.cpr.2020.101861
  • van Eldik, W. M., de Haan, A. D., Parry, L. Q., Davies, P. T., Luijk, M. P. C. M., Arends, L. R., & Prinzie, P. (2020). The interparental relationship: Meta-analytic associations with children's maladjustment and responses to interparental conflict. Psychological Bulletin, 146(7), 553–594. https://doi.org/10.1037/bul0000233
  • World Health Organization. (2026, May 7). Violence against children. WHO fact sheet. https://www.who.int/news-room/fact-sheets/detail/violence-against-children