Abstract
Parental love is routinely invoked in developmental science, clinical practice, family policy, and everyday reasoning, yet it is rarely defined as a measurable psychological construct. This conceptual review argues that the scientifically tractable question is not whether parents "really love" their children, but whether the affective and motivational systems ordinarily denoted by parental love vary continuously across caregivers, relationships, time, and context. We distinguish parental love from observable parenting behavior and from parenting competence, and define it provisionally as a latent disposition comprising affection, valuation of the child, motivation to protect, willingness to incur costs for the child's welfare, and durable commitment to the relationship. Evidence from interpersonal acceptance–rejection research, attachment and sensitivity research, behavioral genetics, neurobiology, intervention science, comparative primatology, evolutionary models of parental investment, and child-maltreatment epidemiology converges on substantial variation in the behavioral and motivational manifestations of this disposition. Crucially, however, maltreatment, insecure attachment, or low sensitivity cannot be treated as direct measures of low love: similar behavior can arise from stress, psychopathology, cultural norms, limited skill, or constrained resources. We therefore propose a multidimensional Parental Affective Investment Model that separates latent affective investment, caregiving competence, contextual constraint, and child-evoked processes. The model yields falsifiable predictions, a measurement agenda, and implications for developmental research and intervention. The central claim is deliberately narrower than a universal-love critique: parental affective investment should be treated as an empirical variable rather than a presumed constant.
1. Introduction: Turning a Cultural Assumption Into an Empirical Question
The proposition that parents love their children is simultaneously plausible, culturally powerful, and scientifically underspecified. In ordinary language, parental love can refer to affection, attachment, protectiveness, sacrifice, loyalty, identification with the child, or a moral commitment to the child's welfare. Developmental research, by contrast, typically measures more specific constructs: warmth, acceptance, sensitivity, responsiveness, involvement, harshness, rejection, monitoring, or investment. The mismatch matters. When a broad folk concept is treated as self-evident, it can enter scientific reasoning as an unmeasured background assumption rather than as a construct whose distribution, determinants, and consequences require evidence.
This article advances a methodological thesis: the affective and motivational phenomena ordinarily grouped under the label parental love should be modeled as variables, not constants. The claim does not require the premise that most parents are unloving, nor does it imply that low-quality parenting is equivalent to low love. Instead, it follows the ordinary logic of individual-differences science. Psychological dispositions that depend on emotion, motivation, learning, physiology, personality, relationship history, and context generally show meaningful between-person and within-person variation. There is little theoretical reason to exempt parental affective investment from that logic.
The existing literature already contains much of the evidence needed to motivate this reframing, but it is fragmented across traditions. Interpersonal acceptance–rejection theory documents cross-cultural variation in perceived parental warmth and rejection and their associations with adjustment . Attachment research shows substantial variation in caregiver sensitivity and in child–caregiver relationship organization, while emphasizing that sensitivity is neither identical to parental affection nor sufficient to determine attachment . Meta-analytic work links observed sensitivity to child behavioral adjustment, although effects are generally modest and reciprocal processes remain important . Behavioral-genetic studies show that measured parenting reflects parent, child, and environmental influences. Intervention research demonstrates that sensitivity and discipline can change, confirming that caregiving behavior is plastic rather than a fixed expression of an immutable trait . (Khaleque & Ali, 2017; Nivison et al., 2026; O'Neill et al., 2021; van IJzendoorn et al., 2023)
The conceptual error to avoid is reverse inference: because parenting behavior varies, parental love must be directly observable in that behavior. It is not. A highly affectionate parent can behave insensitively because of depression, exhaustion, poverty, trauma, limited knowledge, or poor emotion regulation. Conversely, a caregiver can perform competent routines without strong affection. The strongest version of the present thesis is therefore not that maltreatment or insensitivity "proves" absence of love. It is that parental affective investment is a plausible latent construct whose manifestations vary and whose measurement should be separated from competence and constraint.
2. Construct Clarification: What Would "Parental Love" Mean Scientifically?
For scientific purposes, we define parental affective investment as a relatively enduring but context-sensitive disposition to positively value a particular child, experience affiliative affect toward that child, protect the child from harm, incur personal costs for the child's welfare, and maintain commitment to the relationship across time. The term affective investment is preferable to love in formal modeling because it identifies components that can, at least in principle, be measured. "Love" remains useful as the folk construct that motivates the inquiry.
This definition has five implications. First, the construct is relational: the same adult may show different levels or forms of investment toward different children. Second, it is multidimensional: affection, protection, sacrifice, and commitment need not covary perfectly. Third, it is probabilistic rather than deterministic: high investment raises the probability of protective and responsive behavior but does not guarantee it in every situation. Fourth, it can change over time. Fifth, it should not be inferred from a single adverse event, a diagnostic label, or a child outcome.
A central contribution of the proposed framework is the separation of three constructs that are often collapsed. Affective investment concerns motivation and valuation. Caregiving competence concerns the skills required to perceive cues, regulate one's own responses, set developmentally appropriate limits, and provide contingent care. Contextual capacity concerns whether time, money, safety, health, social support, and institutional conditions permit the parent to translate motivation and skill into behavior. The same observed behavior can therefore have different latent causes.
This distinction is supported by intervention evidence. Programs can improve sensitive responsiveness and discipline, indicating that at least part of caregiving quality reflects learnable skill rather than stable affection. A meta-analysis of 25 randomized trials of Video-feedback Intervention to Promote Positive Parenting and Sensitive Discipline synthesized more than 2,000 caregivers and found improvements in parenting-related outcomes. A preregistered study of 120 first-time parents found that a single-session video-based intervention improved observed sensitivity and reduced intrusiveness . Such findings caution against interpreting low baseline skill as low parental attachment or concern. (Bakermans-Kranenburg et al., 2003; Jugovac et al., 2022; van IJzendoorn et al., 2023)
3. Evidence That the Behavioral Manifestations Vary
Interpersonal acceptance–rejection theory provides especially clear evidence that perceived warmth and rejection vary systematically across families and cultures. A systematic review of 12 meta-analyses summarized 551 studies and 149,440 respondents from 31 countries, finding pancultural associations between interpersonal acceptance–rejection and psychological adjustment . These data do not measure "love" directly, but they establish that one of its most obvious experiential manifestations—felt acceptance versus rejection—is neither uniform nor trivial. (Khaleque & Ali, 2017; Rohner, 1980)
The literature also supports a dimensional rather than purely categorical representation. Large meta-analyses have found that warmth and autonomy support were associated with fewer internalizing and externalizing problems, whereas harsh and psychological control were associated with more problems; longitudinal associations were generally small and evidence also supported bidirectionality. The relevant inference is not that warmth causes all downstream outcomes, but that meaningful variability in warmth exists across ordinary populations and participates in reciprocal developmental processes. (Nivison et al., 2026; Pinquart, 2017; Pinquart & Kauser, 2018)
Observed sensitivity is a narrower construct than parental love, but it is valuable because it is behaviorally coded rather than inferred solely from self-report. The most comprehensive recent synthesis is Nivison et al.'s (2026) umbrella review of 17 meta-analyses, drawing on 473 primary studies and 86,625 participants across 35 countries. Across eligible meta-analytic datasets, caregiver sensitivity was moderately associated with attachment security (r = .25), cognition (r = .23), and language (r = .26), but only weakly associated with socioemotional problems (r = -.07). These differentiated effect sizes are theoretically important: sensitivity is a meaningful developmental input, but it is neither a proxy for parental love nor a dominant explanation of child adjustment (Cooke et al., 2022; Deneault et al., 2022; Nivison et al., 2026).
Attachment research reaches a compatible conclusion while warning against construct substitution. In a systematic review and meta-analysis of preschool attachment, caregiver sensitivity was higher among caregivers of secure and organized children, with small-to-medium or medium effects depending on measurement strategy . Attachment security is therefore related to caregiving quality, but it is not a direct assay of parental affection. Child temperament, accumulated interaction history, stress, and other relational processes also contribute. (De Wolff & van IJzendoorn, 1997; O'Neill et al., 2021; Nivison et al., 2026)
4. Neurobiological Perspectives: Anchoring Affective Investment
A significant limitation of purely behavioral and self-report approaches is that affective investment remains an inferred construct with no external referent. Neurobiological research provides convergent evidence that the caregiving motivational system is a measurable, plastic, and individually variable biological system. (Carollo et al., 2025; Swain et al., 2014)
The parental brain comprises a network of regions implicated in reward processing, salience detection, and emotion regulation, including the nucleus accumbens, ventral tegmental area, amygdala, orbitofrontal cortex, and anterior cingulate cortex . Activation in these regions in response to infant stimuli—faces, cries, and odors—indexes the motivational salience of caregiving cues. Importantly, this activation varies systematically across individuals and contexts. Mothers reporting higher levels of caregiving motivation show stronger reward-related activation to infant faces, and baby-schema manipulations that increase perceived cuteness elicit greater nucleus accumbens activity . (Carollo et al., 2025; Glocker et al., 2009; Swain et al., 2014)
Oxytocin is a well-established component of the caregiving biobehavioral system, facilitating mother–infant bonding and modulating responses to infant cues . However, the effects of oxytocin are context-dependent and person-specific rather than uniformly prosocial. This nuance supports the dimensional view: the caregiving system is not simply "on" or "off," but is modulated by hormonal state, early experience, and current context. (Swain et al., 2014)
Perhaps most relevant to the present argument, neurobiological systems supporting caregiving are modifiable by experience. Maternal cortisol levels are associated with reduced brain activation to infant cries and more intrusive parenting behavior, suggesting that stress physiology can downregulate the caregiving motivational system. Conversely, caregiving experience itself shapes neural responses: mothers show structural brain changes in the postpartum period that are associated with positive perceptions of the infant. This plasticity indicates that affective investment is not a fixed trait but a dynamic state that can be up- or downregulated by biological and contextual factors. (Carollo et al., 2025; Swain et al., 2014)
5. Contextual Modulation: Stress, Resources, and Burnout
A distributional model becomes stronger, not weaker, when context is taken seriously. Parenting is enacted under unequal conditions. Economic strain, unsafe environments, work demands, relationship conflict, social isolation, sleep loss, and mental-health symptoms can reduce a caregiver's capacity to behave in accordance with their intentions. This is why adverse behavior cannot be mapped one-to-one onto low affective investment. (Belsky, 1984; Ren et al., 2024; Roskam et al., 2021)
Parental burnout illustrates the point. Burnout is characterized by exhaustion in the parental role, emotional distancing, and a sense of contrast with one's previous parental self; it is associated with neglectful and violent behavior toward offspring (Mikolajczak et al., 2018; Ren et al., 2024; Roskam et al., 2021). Importantly, burnout is modifiable. A recent meta-analysis of 15 studies comprising 18 intervention arms (N = 1,380), including 12 randomized controlled trial arms, found moderate-to-large reductions in parental burnout following psychological and educational interventions, with benefits maintained for up to three months (Urbanowicz et al., 2026). This pattern supports a state-sensitive model in which emotional availability and caregiving behavior can deteriorate and recover as regulatory resources and contextual demands change, without requiring the inference that a stable underlying bond has disappeared.
Accordingly, the proposed model treats observed caregiving as a function of at least four inputs: affective investment, competence, contextual capacity, and child-evoked demands. This prevents a common moral and scientific mistake—using an outcome to infer a single hidden cause when several causal pathways can generate the same behavior.
6. Child Maltreatment: Behavioral Variation, Indirect Evidence About Love
Global maltreatment data make one proposition difficult to dispute: harmful caregiving is not confined to a vanishingly small set of families. UNICEF estimated in 2024 that nearly 400 million children younger than five regularly experience psychological aggression or physical punishment at home. Earlier global meta-analytic work likewise documented substantial prevalence of physical abuse, emotional abuse, sexual abuse, and neglect. These estimates vary by definition, informant, surveillance system, and cultural norm, but the overall burden is large. (Stoltenborgh et al., 2015; UNICEF, 2024)
The inferential significance of these data should be stated carefully. Maltreatment falsifies any claim that parental behavior is uniformly benevolent. It does not, by itself, falsify the claim that many maltreating parents experience some form of affection or attachment. Harsh behavior may coexist with love, distorted beliefs about discipline, substance use, severe stress, psychiatric symptoms, intergenerational learning, or poor regulatory capacity. The empirical target is therefore the joint distribution of affect, motivation, competence, and behavior—not a binary classification of parents as loving or unloving. (Euser et al., 2015; Stoltenborgh et al., 2015)
This distinction also improves clinical utility. Risk assessment should be based on behavior, history, threats, protective capacity, contextual stressors, and validated risk factors rather than on declarations of love. Subjective affection may be genuine yet insufficient to protect a child; conversely, a parent's emotionally muted presentation is not evidence of dangerousness. A dimensional model discourages both romanticization and unwarranted moral inference.
7. Evolutionary and Comparative Perspectives
Evolutionary theory predicts that parental effort is costly and therefore allocated under trade-offs. Trivers' parental-investment framework and subsequent work treat investment as expenditure that benefits offspring at a cost to other components of parental fitness. Contemporary reviews of human paternal care emphasize the flexibility of human caregiving and the importance of ecological, social, and cooperative contexts. This literature is relevant because it makes conditionality theoretically expected: investment should respond to resources, kinship structure, offspring needs, and alternative demands. (Trivers, 1972)
Comparative evidence adds a second caution against human exceptionalism. Nonhuman primates show stable and context-sensitive individual differences in maternal contact, rejection, protection, and responsiveness, with neuroendocrine and experiential correlates. Cross-fostering work in rhesus macaques suggests that abusive maternal behavior can be transmitted through early experience rather than simple genetic inheritance. Comparative findings cannot establish the phenomenology of human love, but they support the broader proposition that caregiving systems are variable, plastic, and biologically embedded. (Maestripieri, 2005; Saltzman & Maestripieri, 2011)
8. Developmental Consequences: Why Variation Matters
Variation in caregiving is consequential even when its latent motivational origins are uncertain. Meta-analytic evidence links lower sensitivity with greater child behavioral problems, although effect sizes are small and heterogeneous . Large syntheses of parenting dimensions similarly associate low warmth and harsh or psychologically controlling parenting with internalizing and externalizing symptoms. Neurobiological reviews further suggest that severe and chronic maltreatment can alter developmental trajectories in neural systems implicated in threat processing, emotion regulation, and reward. (Cooke et al., 2022; Nivison et al., 2026; Pinquart, 2017)
These findings should not be interpreted as showing a simple dose-response relation between "love" and brain development. Parenting is one exposure within a wider developmental ecology, and neurobiological findings are sensitive to adversity type, timing, chronicity, co-occurring deprivation, and genetic liability. A stronger theoretical claim is that caregiving quality constitutes a variable developmental input. The child's nervous system develops in relation to patterns of safety, predictability, responsiveness, threat, and deprivation—not in relation to a culturally presumed constant. (Teicher & Samson, 2016; Teicher et al., 2016)
9. The Parental Affective Investment Model
We propose a four-component Parental Affective Investment Model (PAIM). Component 1, affective investment, includes affiliative affect, valuation of the child, protective motivation, willingness to incur costs, and durable commitment. Component 2, caregiving competence, includes cue detection, perspective taking, emotion regulation, developmental knowledge, and effective limit setting. Component 3, contextual capacity, includes material resources, time, social support, physical and mental health, relationship stability, and environmental safety. Component 4, child-evoked processes, includes child characteristics that alter caregiving demands or elicit differential responses.
Observed parenting behavior is modeled as an output of interactions among these components rather than as a transparent readout of any one component. The model predicts equifinality: similar harsh or insensitive behavior can emerge from low investment, low skill, low capacity, high stress, or combinations thereof. It also predicts multifinality: similarly high affective investment can produce different behavioral outcomes depending on competence and context.
The model is explicitly dimensional. It does not assume a normal statistical distribution, and empirical work should test distributional form rather than impose it. Affective investment may be skewed toward high values, multimodal in some populations, or characterized by ceiling effects on self-report. The essential claim is continuous meaningful variation, not Gaussianity.
The model is also relationship-specific. Within-parent comparisons across siblings, biological and nonbiological children, developmental stages, and periods of stress can help separate stable parental dispositions from dyad-specific and contextual effects. This proposition now has substantial within-family empirical grounding. Jensen and Thomsen's (2024) multilevel meta-analysis synthesized 2,890 effect sizes from 88 sources, representing 26,451 participants across 43 samples; less-favored treatment and greater sibling differences in parenting were associated with more internalizing and externalizing problems. Eradus et al. (2024), synthesizing 215 effect sizes from 13 samples, similarly found a small but significant association between relative differential treatment and sibling differences in behavior problems, with stronger associations for differential hostility than warmth. These findings establish consequential dyad-specific variation in parenting, although they do not identify its motivational cause. Generalizability also requires caution: Jensen and Thomsen's pooled participants were predominantly White and more than half of the samples were from the United States.
10. A Measurement and Research Agenda
A serious science of parental affective investment requires measurement that does not simply rename existing parenting scales. At least five methodological priorities follow.
First, construct development should begin with multimethod measurement. Self-report can assess affection, valuation, protective motivation, commitment, and willingness to sacrifice, but should be paired with child reports, partner reports, behavioral observation, ecological momentary assessment, and consequential choice tasks. Convergence and divergence among methods are theoretically informative rather than nuisances.
Second, studies should establish discriminant validity from warmth, attachment style, parental identity, empathy, conscientiousness, depression, stress, and social desirability. If a proposed love scale is indistinguishable from warmth or from a socially desirable self-description, the new construct adds little.
Third, researchers should test measurement invariance across mothers and fathers, biological and adoptive parents, cultures, socioeconomic contexts, child ages, and family structures. A universal theory cannot be built from a measure whose meaning changes across groups.
Fourth, genetically informative and longitudinal designs should distinguish parent effects, child-evoked effects, and reciprocal transactions. Sibling-comparison and within-family designs are particularly valuable because they test whether the same parent invests differently across children while holding many parent-level characteristics constant. (Avinun & Knafo, 2014; Klahr & Burt, 2014; Klahr et al., 2013)
Fifth, preregistered studies should test incremental prediction. Affective-investment measures should predict protective behavior, persistence of caregiving under cost, repair after conflict, and longitudinal child outcomes above established predictors such as warmth, sensitivity, psychopathology, stress, and socioeconomic conditions. Failure to show incremental validity would be theoretically informative and could indicate that "parental love" is better understood as a folk umbrella for already measured constructs.
11. Implications for Intervention, Child Protection, and Family Research
The model has practical implications, but these should be framed cautiously. For intervention, separating motivation from competence and capacity clarifies treatment targets. Meta-analytic evidence indicates that parenting behavior is modifiable: attachment- and emotion-focused parenting interventions show small effects on externalizing symptoms and small-to-moderate effects on internalizing symptoms relative to waitlist controls (Jugovac et al., 2022), while video-feedback interventions improve parenting-related outcomes and attachment-relevant processes (van IJzendoorn et al., 2023). More recent evidence indicates that parental burnout itself can also be reduced through psychological and educational interventions (Urbanowicz et al., 2026). None of these intervention effects demonstrates that parental affective investment changed. Rather, they show why PAIM must distinguish changes in skill, regulation, and contextual capacity from changes in child-specific valuation or protective motivation.
For child protection, the framework reinforces a behavior-first principle. Decisions should rest on evidence of safety, harm, protective capacity, and validated risk indicators rather than on cultural assumptions that biological parents are necessarily protective or on equally simplistic assumptions that a harmful act proves emotional indifference. The framework is compatible with procedural protections for families while insisting that parental status is not itself an empirical measure of caregiving quality.
For developmental research, the model encourages investigators to treat parental affective investment as a hypothesis requiring operationalization. Doing so could connect literatures that currently speak past one another: warmth and rejection, attachment and sensitivity, parental investment, burnout, behavioral genetics, and maltreatment. The benefit is not a new label but a clearer causal architecture. (Buckels et al., 2015; Hofer et al., 2018; Nivison et al., 2026; Schaller, 2018)
11. Theoretical Novelty: What PAIM Adds
The Parental Affective Investment Model (PAIM) is intended to solve a construct-identification problem rather than to rename good parenting. Its distinctive claim is that the motivational valuation of a particular child should be modeled separately from the skills and circumstances that determine whether motivation becomes effective care. Existing constructs capture important parts of this system, but they do not fully separate child-specific valuation and protective motivation from behavioral competence and contextual opportunity. The parental-care motivational literature provides an especially important bridge: Buckels et al. (2015) validated reliable individual differences in parental care and tenderness, Hofer et al. (2018) distinguished nurturance from protection, and Schaller (2018) synthesized evidence for a broader parental-care motivational system. PAIM therefore does not claim to discover parental motivation. Its proposed novelty is narrower: to model child-specific affective investment jointly with competence, capacity, and child-evoked processes, and to require incremental validity beyond these neighboring constructs.
This distinction imposes a stringent criterion for theoretical usefulness. Parental affective investment should not be retained as a separate construct merely because the phrase parental love is intuitively compelling. It must demonstrate discriminant validity from warmth, sensitivity, empathy, parental identity, attachment representations, stress, depression, and socially desirable responding, and it must provide incremental prediction of consequential caregiving choices. If it fails those tests, the scientifically appropriate conclusion would be that parental love is better treated as a folk umbrella for already established constructs. PAIM is therefore explicitly falsifiable rather than definitionally protected.
A second contribution is the model's multilevel structure. Variation may occur between parents, between children within the same parent, within a parent across developmental periods, and within shorter time scales as stress and resources fluctuate. Recent differential-parenting meta-analyses make the within-family level especially important. Jensen and Thomsen (2024) synthesized 2,890 effect sizes from 88 sources and 43 samples (N = 26,451), while Eradus et al. (2024) synthesized 215 effect sizes from 13 samples. Both found that relative differences in parental treatment covary with sibling differences in adjustment, with differential hostility particularly relevant to externalizing problems. These studies do not measure love or affective investment directly. Their theoretical value is narrower but substantial: they demonstrate that consequential variation occurs at the dyadic level even when many parent-level and family-level characteristics are shared.
Behavioral-genetic evidence reinforces this transactional interpretation. Klahr and Burt (2014) found contributions from parental genetic characteristics, child genetic characteristics through evocative gene-environment correlation, and shared and nonshared environments. Avinun and Knafo (2014) likewise showed that genetically influenced child characteristics account for a meaningful portion of variance in measured parenting. These findings do not establish a heritability of love. They show instead that parenting is partly co-produced by parent and child characteristics and therefore cannot be interpreted as a one-way behavioral expression of a fixed parental trait. (Avinun & Knafo, 2014; Klahr & Burt, 2014; Klahr et al., 2013)
12. Formal Propositions and Falsifiable Predictions
Proposition 1: Dimensionality without assumed normality. Child-specific affective investment should show reliable meaningful variation, but its population distribution should be estimated rather than assumed to be Gaussian. Ceiling effects, skew, mixtures, or multidimensional profiles remain plausible.
Proposition 2: Partial separability. Affective investment should correlate positively with warmth and sensitivity while remaining psychometrically distinguishable from caregiving competence, contextual capacity, parental stress, and social desirability.
Proposition 3: Within-parent differentiation. Repeated assessments across siblings should reveal reliable dyad-specific variance after parent-level characteristics are controlled. Near-zero dyad-specific variance would weaken the relationship-specific component of PAIM.
Proposition 4: Motivation × capacity interaction. High affective investment should predict competent protective behavior most strongly when skill and contextual capacity are adequate. Under severe burnout, illness, material constraint, or threat, correspondence between investment and observed behavior should weaken.
Proposition 5: Incremental validity. Multimethod PAI scores should predict costly protective choices, persistence under burden, repair after conflict, and longitudinal relationship quality beyond warmth, sensitivity, personality, psychopathology, stress, and socioeconomic conditions.
Proposition 6: Differential intervention response. Skill-focused interventions should primarily improve competence; resource or mental-health interventions should improve behavior when contextual capacity is limiting. Failure to distinguish these pathways would count against PAIM's mechanistic usefulness.
Proposition 7: Reciprocal dynamics. Child behavior and developmental stage should alter parental responses and may, over time, alter affective investment itself. Longitudinal transactional models should therefore outperform strictly parent-to-child causal models.
Proposition 8: Convergent but nonredundant biology. Neural, autonomic, endocrine, and behavioral indices of child-specific salience may provide convergent validity, but no biological marker should be treated as a direct or sufficient assay of love.
13. Measurement Architecture for a Cumulative Science
A serious measurement program should avoid a face-valid scale that simply asks parents whether they love their children. Such an instrument would be highly vulnerable to ceiling effects, impression management, and cultural scripts. Construct development should instead use a multitrait-multimethod design. Candidate indicators include child-specific affiliative affect, valuation, protective motivation, willingness to incur meaningful costs, persistence under burden, and commitment to relationship repair. Methods should combine parent report, child and partner report, structured observation, ecological momentary assessment, consequential choice tasks, and, where theoretically justified, physiological or neural responses to child-specific cues.
Existing parental-care measures provide a starting point rather than a finished instrument. The PCAT program demonstrates that nurturance and protection can be measured reliably (Buckels et al., 2015; Hofer et al., 2018), while attachment and sensitivity research supplies mature observational traditions. The next step is child-specific measurement with explicit tests of convergent and discriminant validity. Confirmatory models should compare a higher-order PAI model with correlated-component and bifactor alternatives, and measurement invariance should be tested across mothers and fathers, biological, adoptive, foster, and stepparents, cultural settings, child ages, and family structures. (Buckels et al., 2015; Hofer et al., 2018; Schaller, 2018)
Longitudinal measurement is essential because PAIM distinguishes relatively stable disposition from state modulation. Parental burnout offers a strong natural test. Burnout includes exhaustion and emotional distancing and has been associated with neglect and violence (Mikolajczak et al., 2018); a 42-country study of 17,409 parents found substantial cross-national variation in burnout (Roskam et al., 2021). These findings show why deteriorating caregiving cannot automatically be interpreted as low enduring investment: contextual and regulatory depletion can change behavior and emotional availability over time. (Mikolajczak et al., 2018; Ren et al., 2024; Roskam et al., 2021; Urbanowicz et al., 2026)
The construct should ultimately earn its place through out-of-sample prediction. Preregistered studies should test whether PAI improves prediction of costly caregiving decisions, persistence, safety behavior, and relationship repair above established predictors. Cross-validation, specification-curve analyses, and multiverse analyses would reduce researcher degrees of freedom. Null incremental validity should be treated as a theoretically meaningful result rather than a measurement failure to be explained away.
Figure 1. The Parental Affective Investment Model (PAIM)
Note. Observed caregiving is modeled as an emergent product of affective investment, competence, contextual capacity, and child-evoked processes. Dashed feedback indicates that child characteristics and accumulated interactions can alter subsequent investment and caregiving.
14. Alternative Explanations and Boundary Conditions
Several rival explanations must be taken seriously. Apparent variation in PAI may be reducible to personality, empathy, attachment representations, or generalized caregiving motivation; apparent child-specific variation may primarily reflect child-evoked effects; self-reported investment may be dominated by social desirability and identity maintenance; and culturally patterned expressions of affection may create spurious group differences if Western expressive warmth is treated as the criterion. High investment is also not synonymous with beneficial parenting: overprotection, enmeshment, intrusive control, or coercive behavior can be motivated by intense concern. Each alternative generates testable discriminant predictions and therefore strengthens the framework by constraining it.
The framework should remain agnostic about moral worth and legal status. A low score on a future PAI measure would not by itself establish dangerousness, parental unfitness, or a legal conclusion. Child-protection decisions should remain behavior- and safety-centered. Likewise, biological or legal parenthood should not be used as a proxy for protection. PAIM is intended to improve causal understanding and intervention targeting, not to create a psychometric label for adjudicating parental virtue.
Evolutionary and comparative evidence should be used at the appropriate level of inference. Parental-investment theory predicts costly trade-offs and conditional allocation (Trivers, 1972), while comparative primate research demonstrates variability and experience-dependent plasticity in maternal behavior (Maestripieri, 2005; Saltzman & Maestripieri, 2011). Human neuroscience likewise supports a distributed and plastic caregiving system rather than a unitary 'love center.' Carollo et al.'s (2025) bibliometric review mapped 656 publications spanning 1951–2024 and identified converging research clusters involving responsiveness, reward, fatherhood, and maternal health. These literatures make flexible caregiving motivation biologically plausible; they do not establish a neural assay of love or license inference about the motives of an individual parent.
15. A Staged Research Program
A decisive empirical program can proceed in stages. Stage 1 should define culturally transportable indicators of affection, valuation, protection, cost-bearing, and commitment. Stage 2 should establish factor structure, reliability, method variance, and measurement invariance in diverse family forms. Stage 3 should use sibling and repeated-measures designs to partition parent-general, dyad-specific, and time-varying components. Stage 4 should test incremental and predictive validity against warmth, sensitivity, parental-care motivation, stress, psychopathology, and socioeconomic context. Stage 5 should use interventions and natural experiments to test mechanisms: skill training should primarily alter competence, resource relief should primarily alter capacity, and claims that PAI itself changed should require evidence that motivational valuation changed rather than merely behavior.
This program also clarifies what the current evidence does and does not establish. The literature strongly supports continuous variation in many manifestations of caregiving, but it does not yet demonstrate that a single latent variable called parental love has a particular population distribution. PAIM therefore makes the weaker and more defensible claim of meaningful dimensional variation while leaving distributional form, component structure, and temporal stability open to empirical adjudication. That epistemic restraint is central to the framework's scientific value.
16. Evidentiary Standards and Claims of Causality
PAIM is a theory-building synthesis, not a systematic review, and the evidentiary status of its claims should therefore be explicit. Where available, the present argument gives greatest weight to preregistered umbrella reviews, meta-analyses, randomized trials, genetically informative designs, and large longitudinal or within-family studies. Cross-sectional associations, retrospective reports, neuroimaging correlates, and comparative findings are used primarily for construct triangulation and hypothesis generation rather than causal identification. This hierarchy matters because the same empirical association can support very different claims depending on design.
Causal language should consequently be reserved for designs that warrant it. Meta-analytic associations between sensitivity and child outcomes establish robust covariation but do not, by themselves, identify unidirectional parent-to-child effects; reciprocal and genetically confounded pathways remain plausible. Intervention studies offer stronger leverage on modifiable caregiving processes, but improvements in sensitivity, discipline, or burnout cannot be interpreted as changes in affective investment unless the motivational construct itself is independently measured. The decisive tests of PAIM will therefore require longitudinal, within-family, genetically informed, and experimental designs that measure investment, competence, capacity, and behavior separately.
17. Limitations and Boundary Conditions
The proposed framework has important limitations. Most existing evidence concerns parenting behavior, caregiver sensitivity, perceived acceptance, or generalized parental-care motivation rather than parental affective investment as independently measured. The model is therefore a conceptual synthesis and research program, not evidence that a single latent trait has already been identified. Second, cultural coverage remains uneven. Even large cross-national syntheses do not guarantee measurement equivalence, and some especially relevant within-family evidence remains demographically concentrated; for example, the Jensen and Thomsen (2024) meta-analysis was predominantly White and heavily U.S.-based. Third, behavioral-genetic estimates are population- and design-specific and should not be interpreted as immutable properties of traits (Avinun & Knafo, 2014; Klahr & Burt, 2014). Fourth, neurobiological correlates of caregiving motivation are neither necessary nor sufficient indicators of subjective love (Carollo et al., 2025; Swain et al., 2014). Fifth, evolutionary hypotheses operate at population and functional levels and cannot justify inference about an individual parent's motives. Sixth, child maltreatment represents an extreme behavioral outcome and should not be used to infer the shape of the full population distribution. Finally, because sensitivity predicts some developmental domains more strongly than others, the theory should resist treating any single parenting dimension as a global developmental cause (Nivison et al., 2026).
A further limitation is normative. High parental investment is not synonymous with optimal parenting. Overprotection, enmeshment, intrusive control, or coercive behavior can be motivated by intense concern. Conversely, culturally restrained emotional expression can coexist with deep commitment and effective care. Any valid measure must therefore avoid equating a particular Western expressive style with love itself. (Pinquart, 2017; Pinquart & Kauser, 2018)
18. Conclusion
The most defensible scientific challenge to the idealized-parent assumption is not the claim that parental love is absent in a substantial fraction of parents. Existing evidence cannot support that precise statement because parental love has rarely been operationalized independently of behavior. The stronger contribution is methodological: parental affective investment should not be presumed constant merely because parenthood is biologically or socially salient.
Across acceptance–rejection research, observed sensitivity, behavioral genetics, neurobiology, intervention science, comparative primatology, evolutionary theory, and maltreatment epidemiology, caregiving-related affect and behavior show systematic variation. The task for the next generation of research is to determine how much of that variation reflects differences in affective investment, how much reflects competence and constraint, how much is evoked by the child, and how these components interact over time.
Replacing presumption with measurement avoids two symmetrical errors. It avoids romanticizing parenthood by assuming that biological or legal status guarantees uniformly protective motivation. It also avoids moral reductionism by assuming that insensitive or harmful behavior transparently reveals absence of love. A multidimensional, testable model is more scientifically useful than either extreme. The central proposition is therefore modest but consequential: parental love, if it is to function as a scientific construct at all, must be measured as a variable.
Keywords: parental love, parental warmth, parental sensitivity, parental investment, acceptance–rejection, parenting, behavioral genetics, attachment, child maltreatment, individual differences
References
Ainsworth, M. D. S., Blehar, M. C., Waters, E., & Wall, S. (1978). Patterns of attachment: A psychological study of the Strange Situation. Lawrence Erlbaum Associates.
Avinun, R., & Knafo, A. (2014). Parenting as a reaction evoked by children's genotype: A meta-analysis of children-as-twins studies. Personality and Social Psychology Review, 18(1), 87–102. https://doi.org/10.1177/1088868313498308
Bakermans-Kranenburg, M. J., van IJzendoorn, M. H., & Juffer, F. (2003). Less is more: Meta-analyses of sensitivity and attachment interventions in early childhood. Psychological Bulletin, 129(2), 195–215. https://doi.org/10.1037/0033-2909.129.2.195
Belsky, J. (1984). The determinants of parenting: A process model. Child Development, 55(1), 83–96. https://doi.org/10.2307/1129836
Buckels, E. E., Beall, A. T., Hofer, M. K., Lin, E. Y., Zhou, Z., & Schaller, M. (2015). Individual differences in activation of the parental care motivational system: Assessment, prediction, and implications. Journal of Personality and Social Psychology, 108(3), 497–514. https://doi.org/10.1037/pspp0000023
Carollo, A., Torre, L., Bornstein, M. H., & Esposito, G. (2025). The parental brain: Anatomization of 75 years of neuroscience 1951–2024. Neuroscience Research, 216, 104890. https://doi.org/10.1016/j.neures.2025.03.002
Cooke, J. E., Deneault, A.-A., Devereux, C., Eirich, R., Fearon, R. M. P., & Madigan, S. (2022). Parental sensitivity and child behavioral problems: A meta-analytic review. Child Development, 93(5), 1231–1248. https://doi.org/10.1111/cdev.13764
De Wolff, M. S., & van IJzendoorn, M. H. (1997). Sensitivity and attachment: A meta-analysis on parental antecedents of infant attachment. Child Development, 68(4), 571–591.
Deneault, A.-A., Cabrera, N. J., & Bureau, J.-F. (2022). A meta-analysis on observed paternal and maternal sensitivity. Child Development, 93(6), 1631–1648. https://doi.org/10.1111/cdev.13832
Eradus, M., Leijten, P., Melendez-Torres, G. J., Foo, X. Q., & Oliver, B. R. (2024). Parental differential warmth, hostility, and sibling differences in internalizing and externalizing behavior problems: A meta-analysis. Journal of Family Psychology, 38(3), 387–399. https://doi.org/10.1037/fam0001194
Euser, S., Alink, L. R. A., Stoltenborgh, M., Bakermans-Kranenburg, M. J., & van IJzendoorn, M. H. (2015). A gloomy picture: A meta-analysis of randomized controlled trials reveals disappointing effectiveness of programs aiming at preventing child maltreatment. BMC Public Health, 15, 1068. https://doi.org/10.1186/s12889-015-2387-9
Glocker, M. L., Langleben, D. D., Ruparel, K., Loughead, J. W., Valdez, J. N., Griffin, M. D., Sachser, N., & Gur, R. C. (2009). Baby schema modulates the brain reward system in nulliparous women. Proceedings of the National Academy of Sciences, 106(22), 9115–9119. https://doi.org/10.1073/pnas.0811620106
Hofer, M. K., Buckels, E. E., White, C. J. M., Beall, A. T., & Schaller, M. (2018). Individual differences in activation of the parental care motivational system: An empirical distinction between protection and nurturance. Social Psychological and Personality Science, 9(8), 907–916. https://doi.org/10.1177/1948550617728994
Jensen, A. C., & Thomsen, A. E. (2024). Parental differential treatment of siblings linked with internalizing and externalizing behavior: A meta-analysis. Child Development, 95(4), 1384–1405. https://doi.org/10.1111/cdev.14091
Jugovac, S., O'Kearney, R., Hawes, D. J., & Pasalich, D. S. (2022). Attachment- and emotion-focused parenting interventions for child and adolescent externalizing and internalizing behaviors: A meta-analysis. Clinical Child and Family Psychology Review, 25(4), 754–773. https://doi.org/10.1007/s10567-022-00401-8
Khaleque, A., & Ali, S. (2017). A systematic review of meta-analyses of research on interpersonal acceptance–rejection theory: Constructs and measures. Journal of Family Theory & Review, 9(4), 441–458.
Klahr, A. M., & Burt, S. A. (2014). Elucidating the etiology of individual differences in parenting: A meta-analysis of behavioral genetic research. Psychological Bulletin, 140(2), 544–586. https://doi.org/10.1037/a0034205
Klahr, A. M., Thomas, K. M., Hopwood, C. J., Klump, K. L., & Burt, S. A. (2013). Evocative gene–environment correlation in the mother–child relationship: A twin study of interpersonal processes. Development and Psychopathology, 25(1), 105–118. https://doi.org/10.1017/S0954579412000934
Maestripieri, D. (2005). Early experience affects the intergenerational transmission of infant abuse in rhesus monkeys. Proceedings of the National Academy of Sciences, 102(27), 9726–9729. https://doi.org/10.1073/pnas.0504122102
Mikolajczak, M., Brianda, M. E., Avalosse, H., & Roskam, I. (2018). Consequences of parental burnout: Its specific effect on child neglect and violence. Child Abuse & Neglect, 80, 134–145. https://doi.org/10.1016/j.chiabu.2018.03.025
Nivison, M. D., Fearon, P., Jenkins, J. M., & Madigan, S. (2026). Annual Research Review: The role of caregiver sensitivity in children's developmental outcomes—An umbrella review. Journal of Child Psychology and Psychiatry, 67(4), 486–507. https://doi.org/10.1111/jcpp.70087
O'Neill, M. C., Badovinac, S., Pillai Riddell, R., Bureau, J.-F., Rumeo, C., & Costa, S. (2021). The longitudinal and concurrent relationship between caregiver sensitivity and preschool attachment: A systematic review and meta-analysis. PLOS ONE, 16(1), e0245061. https://doi.org/10.1371/journal.pone.0245061
Piersiak, H. A., Takemoto, S., Bailes, L. G., Kwasneski, K., Salo, V. C., King, L. S., & Humphreys, K. L. (2025). Enhancing caregiver sensitivity: Efficacy of a single-session, video-based training for new parents. Developmental Psychology. Advance online publication.
Pinquart, M. (2017). Associations of parenting dimensions and styles with externalizing problems of children and adolescents: An updated meta-analysis. Developmental Psychology, 53(5), 873–932. https://doi.org/10.1037/dev0000295
Pinquart, M., & Kauser, R. (2018). Do the associations of parenting styles with behavior problems and academic achievement vary by culture? Results from a meta-analysis. Cultural Diversity and Ethnic Minority Psychology, 24(1), 75–100. https://doi.org/10.1037/cdp0000149
Ren, X., Cai, Y., Wang, J., & Chen, O. (2024). A systematic review of parental burnout and related factors among parents. BMC Public Health, 24, 376. https://doi.org/10.1186/s12889-024-17829-y
Rohner, R. P. (1980). Antecedents and consequences of parental rejection: A theory of emotional abuse. Child Abuse & Neglect, 4(3), 189–198.
Roskam, I., et al. (2021). Parental burnout around the globe: A 42-country study. Affective Science, 2, 58–79. https://doi.org/10.1007/s42761-020-00028-4
Saltzman, W., & Maestripieri, D. (2011). The neuroendocrinology of primate maternal behavior. Progress in Neuro-Psychopharmacology and Biological Psychiatry, 35(5), 1192–1204.
Schaller, M. (2018). The parental care motivational system and why it matters (for everyone). Current Directions in Psychological Science, 27(5), 295–301. https://doi.org/10.1177/0963721418767873
Stoltenborgh, M., Bakermans-Kranenburg, M. J., Alink, L. R. A., & van IJzendoorn, M. H. (2015). The prevalence of child maltreatment across the globe: Review of a series of meta-analyses. Child Abuse Review, 24(1), 37–50.
Swain, J. E., Kim, P., Spicer, J., Ho, S. S., Dayton, C. J., Elmadih, A., & Abel, K. M. (2014). Approaching the biology of human parental attachment: Brain imaging, oxytocin and coordinated assessments of mothers and fathers. Brain Research, 1580, 78–101.
Teicher, M. H., & Samson, J. A. (2016). Annual Research Review: Enduring neurobiological effects of childhood abuse and neglect. Journal of Child Psychology and Psychiatry, 57(3), 241–266. https://doi.org/10.1111/jcpp.12507
Teicher, M. H., Samson, J. A., Anderson, C. M., & Ohashi, K. (2016). The effects of childhood maltreatment on brain structure, function and connectivity. Nature Reviews Neuroscience, 17(10), 652–666. https://doi.org/10.1038/nrn.2016.111
Trivers, R. L. (1972). Parental investment and sexual selection. In B. Campbell (Ed.), Sexual selection and the descent of man, 1871–1971 (pp. 136–179). Aldine.
UNICEF. (2024, June 11). Nearly 400 million young children regularly experience violent discipline at home.
Urbanowicz, A. M., Verger, N. B., Shankland, R., Rance, J., Bennett, P., & Gauchet, A. (2026). A meta-analysis of parental burnout interventions. Journal of Affective Disorders, 399, 121022. https://doi.org/10.1016/j.jad.2025.121022
van IJzendoorn, M. H., Schuengel, C., Wang, Q., & Bakermans-Kranenburg, M. J. (2023). Improving parenting, child attachment, and externalizing behaviors: Meta-analysis of the first 25 randomized controlled trials on the effects of Video-feedback Intervention to promote Positive Parenting and Sensitive Discipline. Development and Psychopathology, 35(1), 241–256. https://doi.org/10.1017/S0954579421001462