Adolescent Affective Psychopathic Traits and Infant Attachment Security: A 14-Year Longitudinal Study
Theoretical Background and Conceptual Framework
Adolescent affective psychopathic traits represent the callous–unemotional dimension of psychopathy, characterized fundamentally by a persistent deficiency in empathy, guilt, and remorse.
These traits serve as robust predictors of lagged aggression and severe violence in adolescence, even after statistically controlling for prior violent behavior, individual self-control, peer delinquent influences, and parenting practices.
Etiological models of affective psychopathic traits incorporate multiple distinct domains:
Biological and neurobiological risk factors: dysfunction within neural circuits involving the amygdala and frontal cortex, alongside substantial genetic heritability.
Environmental and family adversities: early exposure to socioeconomic hardship, maternal depressive symptomatology, elevated parental distress, and harsh or hostile parenting practices.
Protective environmental factors: consistent parental warmth, which demonstrates strong inverse associations with callous–unemotional behaviors.
John Bowlby's attachment theory provides a core developmental framework conceptualizing how early relational experiences organize lifelong socio-emotional functioning:
Infants construct an "internal working model"—a cognitive representational schema derived from early caregiver interactions that directs how an individual perceives, interprets, and responds to social environments and relationships.
Bowlby's historical investigation of 44 juvenile thieves established an early clinical connection between maternal separation prior to age 5 and the emergence of severe emotional blunting and remorselessness, termed "affectionless psychopathy."
Multiple developmental pathways link early attachment insecurity to subsequent psychopathology:
Parenting feedback loops: Insecure infant behaviors evoke suboptimal caregiving responses; mothers of insecurely attached infants frequently exhibit lower empathic understanding and systematically misperceive infant mental and emotional states.
Emotion regulation capacities: Dyadic interactions between infant and primary caregiver form the biological and psychological template for developing autonomous affective regulation.
Social-cognitive representations: Severe separations, caregiver unavailability, and attachment disruptions create maladaptive internal working models characterized by hostile attributions, social mistrust, and diminished expectations of interpersonal warmth, fostering callous affective styles.
Classification and Behavioral Dynamics of Infant Attachment Styles
Individual differences in infant attachment security are evaluated through Mary Ainsworth's Strange Situation Procedure (SSP), an eight-episode laboratory paradigm that introduces mild, cumulative stress through alternating separations and reunions with the caregiver and exposure to an unfamiliar adult.
Attachment styles are classified based on the infant's behavioral balance between environmental exploration and proximity seeking, with primary emphasis on reunion behaviors:
Secure Attachment ():
Behavioral pattern: Infants demonstrate a clear preference for the primary caregiver over a stranger, utilize the caregiver as a secure base from which to explore the environment, show explicit signs of missing the parent during separation, and actively seek proximity and physical contact upon reunion.
Regulatory outcome: Distress is easily soothed by the caregiver, allowing rapid recovery and return to exploratory play.
Insecure–Avoidant Attachment ():
Behavioral pattern: Infants show minimal outward distress during separation, exhibit little differentiation between the caregiver and a stranger, and actively avoid, ignore, or move away from the caregiver upon reunion.
Regulatory dynamic: The infant suppresses emotional distress and minimizes proximity-seeking behaviors to avoid parental rejection, leading to a long-term strategy of emotional detachment, defensive independence, and social alienation. This defensive inhibition impedes the capacity to cultivate empathy or experience moral remorse.
Insecure–Resistant/Ambivalent Attachment ():
Behavioral pattern: Infants exhibit pervasive anxiety, anger, and extreme distress upon separation, combined with persistent proximity-seeking that is fused with overt resistance, petulance, or angry rejection upon the caregiver's return.
Regulatory dynamic: Infants remain chronically dysregulated and cannot be comforted, reflecting an underlying strategy of hyperactivating attachment behaviors to maintain engagement with an inconsistently responsive caregiver.
Insecure–Disorganized Attachment ():
Behavioral pattern: Infants display contradictory, disoriented, or fragmented behaviors in the caregiver's presence (e.g., freezing, apprehension, turning in circles, or approaching while averting gaze).
Regulatory dynamic: The caregiver represents simultaneously the source of fear and the biological haven of safety, resulting in a breakdown of an organized behavioral coping strategy.
Contrast with externalizing psychopathology:
Meta-analytic research encompassing 35 studies and 3,568 children indicates that avoidant () and disorganized () attachment classifications significantly increase the risk for externalizing problems (aggression, conduct problems, oppositional behavior, and hostility).
Insecure–resistant attachment () does not predict elevated externalizing conduct problems.
While externalizing problems and affective psychopathic traits both share disinhibitory features, their cognitive-affective profiles are distinct: psychopathy is defined by deficient empathy and hypo-reactivity, whereas general externalizing problems often stem from emotional hyper-reactivity.
Longitudinal Study Methodology and Measurement Protocols
Study Cohort and Sampling Framework:
Longitudinal data were drawn from the National Institute of Child Health and Human Development (NICHD) Study of Early Child Care and Youth Development (SECCYD).
The full SECCYD recruited families across 10 distinct research sites throughout the United States, tracking children from 1 month postpartum through age 15.
Analytic sub-sample selection:
Total baseline sample: .
Excluded due to missing Strange Situation Procedure assessments at 15 months: .
Excluded due to unclassifiable Strange Situation video recordings: .
Final analytic sample: mother–infant dyads.
Sociodemographic profile of the sample:
Child sex: 50.57% male (), 49.43% female ().
Racial composition: 81.46% White (), 11.92% Black (), 1.65% Asian or Pacific Islander (), 0.35% American Indian, Eskimo, or Aleutian (), and 4.61% other backgrounds ().
Child birth order: 44.13% firstborn (), 35.51% second-born (), 15.06% third-born (), and 4.26% fourth-born or later ().
Parental educational attainment at birth: 37.25% of mothers and 38.12% of fathers attained a bachelor's degree or higher.
Household socioeconomic status: 48.22% had an income-to-needs ratio exceeding (middle-to-upper class), 19.93% had ratios below (low income), and 12.71% had ratios below (poverty threshold).
Infant Attachment Assessment ():
Administered via the standard Strange Situation Procedure.
Video coding was conducted by a team of three expert coders with a minimum of 4 years of prior experience, trained on master-coded tapes (including tapes coded by Mary Main) and certified via the University of Minnesota Attachment Test Tapes.
Coding process: Double-coded all recordings; disagreements were resolved via team conferencing to reach consensus. The interrater agreement was 83% with a Cohen's kappa of .
Four specific behavioral categories were systematically rated during reunion episodes: proximity seeking, contact maintaining, resistance, and avoidance, alongside overall exploratory behaviors.
Four-way classification ( distribution):
Secure (): ()
Insecure–Avoidant (): ()
Insecure–Resistant (): ()
Insecure–Disorganized (): ()
Forced three-way classification ( distribution; assigning cases into primary classifications):
Secure (): ()
Insecure–Avoidant (): ()
Insecure–Resistant (): ()
Of note, 113 of the 177 disorganized () infants were classified into the secure () category under the forced three-way scheme.
Adolescent Affective Psychopathic Traits ():
Assessed via the adolescent self-report on the affective factor of the Inventory of Youth Psychopathy Traits (YPI).
Measured across 15 total items scored on a 4-point Likert scale ranging from to .
Comprises three subscales of 5 items each, with continuous sum scores ranging from to :
Remorselessness (; internal consistency ): e.g., "To feel guilt and regret when you have done something wrong is a waste of time."
Unemotionality (; internal consistency ): e.g., "What scares others usually doesn't scare me."
Callousness (; internal consistency ): e.g., "When other people have problems, it is often their own fault, therefore, one should not help them."
Early Covariates and Confounding Variables:
Infant Temperament (): Measured via maternal report on 59 items from the Revised Infant Temperament Questionnaire (RITQ) on a 6-point scale ( to ). Composite average across eight behavioral dimensions: physical motion, approach/withdrawal, adaptability, intensity, mood, persistence, rhythmicity, and sensory threshold ( ). Higher values indicate a more difficult temperament.
Sociodemographic Controls: Maternal age at childbirth, infant biological sex, infant race, maternal education (years), paternal education (years), infant birth order, parental marital/cohabitation status at 15 months, and family income-to-needs ratio at 15 months.
Statistical Analytic Strategy and Propensity Score Weighting
Methodological Advantages of Propensity Scores:
Traditional regression adjustments with multiple covariates can suffer from model misspecification, high dimensionality, and inappropriate extrapolation beyond the support of empirical data.
Propensity score approaches compress high-dimensional baseline covariates into a single scalar summary, enabling causal-style inference in non-randomized designs.
Iterative Generalized Boosted Models (GBM):
Multinomial propensity scores for the four attachment categories () were estimated via the
mnpsfunction in thetwangpackage.Generalized boosted models utilize iterative, nonparametric, tree-based machine learning regressions to optimize covariate balance across multiple treatment groups without overfitting.
Inverse Probability of Treatment Weighting (IPTW):
Weights were generated as the inverse of the estimated propensity of belonging to the observed attachment group ().
Weights were capped at a ceiling value of to prevent extreme weights from inflating sample variance and destabilizing estimations.
Missing Data and Multiple Imputation:
Missing data mechanisms were evaluated using Little's Missing Completely at Random (MCAR) test: , , confirming data were not missing completely at random.
Missingness on adolescent psychopathy measures was significantly related to parental sociodemographics:
Maternal age: ,
Maternal education: ,
Paternal education: ,
Older and more educated parents were more likely to provide complete 15-year adolescent data.
Infant attachment classification was not associated with missingness at 15 years: , .
Multiple imputation was executed using the
micepackage in R, generating 500 imputed datasets via Predictive Mean Matching (PMM) to accommodate non-normal distributions.Final parameters across the 500 imputed sets were pooled using Rubin's rules.
Empirical Findings and Quantitative Results
Bivariate Relationships and Descriptive Statistics:
Callousness at 15 years: Mean , Standard Deviation , Missing .
Remorselessness at 15 years: Mean , Standard Deviation , Missing .
Unemotionality at 15 years: Mean , Standard Deviation , Missing .
Maternal age at childbirth: Mean , Standard Deviation , Missing .
Maternal years of education: Mean , Standard Deviation , Missing .
Paternal years of education: Mean , Standard Deviation , Missing .
Infant temperament composite: Mean , Standard Deviation , Missing .
Income-to-needs ratio at 15 months: Mean , Standard Deviation , Missing .
Intercorrelations among adolescent traits:
Callousness and Remorselessness: ,
Callousness and Unemotionality: ,
Remorselessness and Unemotionality: ,
Correlations between sociodemographic indicators and adolescent traits:
Maternal age was negatively correlated with callousness (, ), remorselessness (, ), and unemotionality (, ).
Maternal education was negatively correlated with callousness (, ), remorselessness (, ), and unemotionality (, ).
Paternal education was negatively correlated with callousness (, ), remorselessness (, ), and unemotionality (, ).
Income-to-needs ratio was negatively correlated with remorselessness (, ) and unemotionality (, ).
Parental marriage/cohabitation at 15 months was negatively correlated with callousness (, ), remorselessness (, ), and unemotionality (, ).
Male sex was positively correlated with callousness (, ), remorselessness (, ), and unemotionality (, ).
Multiple Linear Regression Analysis of Baseline Covariates:
Biological sex: Adolescent males exhibited significantly higher trait scores across all dimensions:
Callousness: , , ,
Remorselessness: , , ,
Unemotionality: , , ,
Family structural stability: Adolescents whose parents were married or cohabiting at 15 months showed significantly lower levels of remorselessness (, , , ).
Sibling constellation: Having three or more older siblings was associated with significantly elevated risks for adolescent unemotionality (, , , ).
Covariate Balance Achieved by Propensity Score Weighting:
Substantial initial imbalances existed before weighting, notably father's education between secure () and avoidant () infants (Standardized Mean Difference ), maternal age (), and marital status ().
Post-IPTW weighting reduced standardized mean differences across nearly all demographic and temperamental variables to below , ensuring balanced covariate distributions across the attachment groups.

Weighted Generalized Linear Model (GLM) Outcomes (Reference Group: Secure Attachment ):
Callousness at 15 years:
Intercept: , , , ,
Insecure–Avoidant (): , , , , (Statistically significant)
Insecure–Resistant (): , , , ,
Insecure–Disorganized (): , , , ,
Remorselessness at 15 years:
Intercept: , , , ,
Insecure–Avoidant (): , , , , (Statistically significant)
Insecure–Resistant (): , , , , (Statistically significant)
Insecure–Disorganized (): , , , ,
Unemotionality at 15 years:
Intercept: , , , ,
Insecure–Avoidant (): , , , , (Statistically significant)
Insecure–Resistant (): , , , ,
Insecure–Disorganized (): , , , ,
Developmental Pathways, Theoretical Synthesis, and Clinical Implications
Longitudinal Significance of Insecure–Avoidant Attachment:
Insecure–avoidant attachment classified during infancy acts as a specific, long-term developmental risk factor predicting elevated levels across all three dimensions of affective psychopathic traits (callousness, remorselessness, and unemotionality) 14 years later.
The findings converge with empirical data from adult community cohorts (e.g., Australian and U.S. adult samples, mean ages and ), where avoidant attachment in romantic and interpersonal relationships is linked to the affective components of psychopathy, specifically meanness and callousness.
Conceptual explanation: Avoidant infants experience low maternal sensitivity and availability, leading them to adaptively abandon proximity seeking and suppress emotional vulnerability. Over developmental time, this sustained emotional detachment crystallizes into chronic social alienation, severe deficits in empathy, and an inability to experience guilt or remorse.
Developmental Specificity of Insecure–Resistant Attachment:
Resistant attachment predicted increased remorselessness at age 15 (, ), but was unrelated to callousness or unemotionality.
Unlike avoidant individuals who suppress emotion, resistant individuals experience heightened, dysregulated emotionality. Their elevated remorselessness may reflect chronic interpersonal conflict or externalizing attributions of blame rather than an underlying lack of emotionality.
Null Findings Regarding Disorganized Attachment:
Disorganized attachment was not associated with elevated adolescent affective psychopathic traits once sociodemographic characteristics and infant temperament were controlled via propensity score weighting ().
Reconciliation with existing literature:
Prior cross-sectional studies linking disorganized attachment to childhood callous–unemotional traits relied largely on small, clinic-referred cohorts.
In large-scale, low-risk community samples, disorganized attachment has not demonstrated an independent longitudinal contribution to callous–unemotional traits when comprehensive covariates are included (e.g., Wright et al., 2018).
Under forced three-way classification, 64% ( of ) of disorganized infants in this cohort were reclassified as secure, indicating a lower baseline risk profile within community samples relative to clinical populations.
Developmental Psychopathology and Canalization:
Early attachment experiences establish initial internal working models that influence how children interpret subsequent social events.
Through developmental canalization and transactional feedback loops, avoidant behavioral strategies evoke less parental sensitivity over childhood, stabilizing into entrenched psychopathological traits by adolescence.
Clinical and Intervention Implications:
Timely identification of early attachment patterns presents a key window for preventive intervention before affective deficits solidify.
While attachment-informed programs are frequently applied to adolescent internalizing disorders, deliberate applications to callous–unemotional and conduct-disordered youth remain limited.
Prior intervention research demonstrates that youth exhibiting elevated callous–unemotional traits achieve significant reductions in internalizing and externalizing symptoms following attachment-based parental interventions, supporting the clinical utility of targeting parent–child attachment mechanisms.
Methodological Limitations and Directions for Future Research
Unmeasured Mediational Mechanisms:
The 14-year longitudinal design did not directly assess intervening causal mechanisms (e.g., intermediary social-cognitive biases, peer rejection, or ongoing parenting shifts) bridging infant avoidant attachment to adolescent traits, leaving precise mediation pathways to be clarified by future studies.
Exclusion of Paternal Attachment Relationships:
Attachment was evaluated exclusively within the mother–child dyad, omitting the distinct contribution of father–child attachment.
Fathers often influence socio-emotional trajectories via "activation parenting" and challenging physical play, requiring family-systems methodologies that examine coparenting dynamics and multiple attachment figures.
Measurement Reliability Constraints:
Adolescent affective traits were captured using 5-item self-report subscales on the Inventory of Youth Psychopathy Traits, yielding modest internal consistency reliability coefficients ( to ).
Sociodemographic Homogeneity and Generalizability:
The study cohort was disproportionately White (81.46%) and predominantly middle class (48.22% with income-to-needs ratios ), restricting the direct generalizability of the findings to more ethnically diverse, economically marginalized, or high-risk clinical populations.