Quality of attachment, disorganization in attachment, and the contribution of caregiver interactions in the home were investigated for infants prenatally exposed to PCP and cocaine and their caregivers. The drug-exposed infants were compared with infants of similar ethnicity, socioeconomic status (SES), and living in the same geographical area of the city with non-substance-abusing mothers. The majority of drug-exposed infants were insecurely attached to their caregivers and did not differ in the percentage of security in the three caregiving environments in which the infants were growing up: biologic mother care, kinship care, or foster mother care. The majority of drug-exposed children were disorganized. Change in caregivers during the first year was not found to be related to the rate of insecurity in any of the caregiving environments. The majority of the non-drug-exposed comparison infants were securely attached, and only a small percentage were disorganized. The high incidence of insecurity in the drug-exposed group is discussed in relation to maternal and environmental circumstances that can alter the assumption of security in attachment for the majority of children and caregivers toward insecurity in attachment.

We tested the hypothesis that attachment security at 12 months can be predicted from infant response to the mother's still-face interaction during the first half-year. Subjects were 66 primiparous mother-infant pairs drawn from the working- and middle-socioeconomic strata (SES). Half of the mothers had experienced a postpartum depression. Infants were observed longitudinally in the still-face paradigm at 2, 4, and 6 months and in the Strange Situation at 12 months. Positive expressions in response to the still-face at 6 months predicted secure attachment at 12 months. Absence of positive expressions predicted avoidant attachment. In infants of postpartum depressed mothers, hours of nonmaternal care positively correlated with secure attachment. Variables associated with infant proneness to distress and sociability were unrelated to attachment security. These results suggest that attachment formation is underway by 6 months of age and that indicators of 12-month attachment security can be detected as early as 6 months. Similar mechanisms of affect regulation and coping may underlie infant response in the still-face and Strange Situation procedures.

This study of 71 low-income mothers and infants examined whether the disorganized/disoriented (D) infant attachment classification is best viewed as a single category or whether at least two subgroups exist, corresponding to the forced-secure and forced-insecure alternate classifications. Correlates of the D classification as a whole, and of the two subtypes of disorganized behavior, were examined in five domains, including 6-month stability, maternal childhood history of loss, severity of maternal psychosocial risk, maternal behavior toward the infant at home, and infant mental development. Results indicated that the two subtypes of disorganized infant attachment behavior differed in age of emergence, maternal childhood history, severity of associated family risk factors, and the extent of the mother's lack of involvement with the infant at home. Across both D subtypes, disorganization of attachment strategies was associated with less optimal maternal behavior at home and with decreased mental development scores at 18 months. Results are discussed in relation to Main and Hesse's (1990) theory of the role of fear-inducing parental behavior in the genesis of disorganized infant attachment behavior and in relation to Aber, Allen, Carlson, and Cicchetti's (1989) concept of secure readiness to learn.

Attachment theory predicts that maltreated children will form insecure patterns of attachment to their caregivers and that attachment relationships are open to change with development. In this cross-sectional/longitudinal investigation, we examined the attachment patterns of 125 maltreated and nonmaltreated preschoolers from the low socioeconomic strata (SES). Maltreated and demographically matched nonmaltreated comparison children were assessed in the Strange Situation at 30, 36, and 48 months of age, along with a subsample of children who were observed longitudinally across a 6- to 18-month period. Attachment relations were classified using a newly developed system for assessing attachment in the preschool years by Cassidy and Marvin (1991). Results revealed that, at each age, maltreated children were significantly more likely to evidence insecure patterns of attachment to their caregivers. The specific types of insecurity demonstrated by children varied with age. Longitudinally, the high percentage of nonmaltreated children who were classified as securely attached were likely to remain securely attached at subsequent assessments. In contrast, the small number of maltreated children who evidenced secure attachments were unlikely to be classified as secure at later assessments. These data provide new information on the patterns of attachment maltreated children exhibit in the preschool years.

This study attempted to replicate previous findings that indicated that preschool-aged children with significant externalizing problems are more likely to have insecure attachment relations than nonproblem peers (Speltz, Greenberg, & DeKlyen, 1990). Fifty children (ages 3–6) and their mothers participated; 25 referred to a child psychiatry clinic for one of the DSM-IIIR Disruptive Behavior Disorders, and 25 matched comparisons without externalizing behavior problems. Attachment was measured at the time of referral using two separation and reunion sequences in a clinic context. We found that 80% of the clinic-referred children were classified as insecure, whereas only 28% of the comparisons were so classified (p <.001). Further, a high percentage of clinic insecure children showed a controlling pattern of attachment. Thus, the present study, with improved methodological rigor, provided an almost exact replication to previous results. The discussion focuses on what contribution attachment research can provide to a developmental psychopathology perspective on early externalizing disorders.

Knowledge of parental state of mind with respect to parent-child attachment relationships may provide insight into etiology or maintenance of problematic parent-child interactions and development of child psychopathology. Forty-nine mothers and their behaviorally disturbed children, aged 5 to 11 years, were assessed in a child psychiatry clinic. Mothers and children were observed in a semistructured interaction consisting of free play, three tasks, separation, and reunion. Mothers were scored on supportiveness, helpfulness, organization, and so forth. Children were rated on relationship behaviors with mother, task behaviors, activity level, and so forth. Parents and teachers completed rating scales of child aggression and oppositionality, inattention and hyperactivity, and competence. Children rated themselves on feelings of depression and anxiety. Mothers were given the Adult Attachment Interview (George, Kaplan, & Main, 1985) and were classified as secure, dismissing, or preoccupied in state of mind with respect to attachment. Mothers classified as secure were supportive and well organized with their children. Mothers classified as insecure were unsupportive and cool. The secure classification was associated with the children reporting low levels of anxiety and depression, being described as competent and relatively low in symptomatology. The dismissing classification was associated with oppositional and aggressive symptoms in the children, greater symptomatology overall, and child self-reports of distress. Maternal attachment classification appears to contribute to understanding of the development of psychopathology in children and its type and severity.

Continuity in individual patterns of adaptation and maladaptation was examined across the years from infancy to preadolescence. Forty-seven preadolescent subjects from a 15-year longitudinal study of children at risk for developmental problems participated. Subjects were intensively observed in a 4-week summer daycamp at age 10. Children were then rated on scales of agency, social skill, dependency, positive and negative affect, and ranked on emotional health. In addition, Q-sorts were performed to obtain an overall description of the child's personality, and observations were made of child-adult interactions. Significant differences were found between secure and insecure attachment groups on molar measures of adaptation including social competence and ego resiliency. Children with anxious histories were also rated higher on dependency and were found to be more frequently involved in interactions with adults than children with secure histories. In addition, differences were found in the patterns of adaptation exhibited by children with the two types of anxious attachment history. Children with resistant histories were found to be more often the recipients of adult-initiated contact, specifically support and nurturance, than were children with avoidant attachment histories. Comparison of overall personality style through correlations of composite counselor Q-sorts identified children with secure and resistant attachment histories as most like other children with the same attachment history. Results for avoidant children did not achieve significance. Patterns of contemporary child behavior revealed through counselor Q-sorts were also found to be related to theoretically derived predictions from patterns observed in infancy.

This article uses a developmental pathway model to consider the role of attachment processes in adolescents' reports of depressive symptoms. Teen attachments were assessed with both interview and observational methods. Assessments of teens' strategies in the Adult Attachment Interview (AAI) indicated that insecure and preoccupied strategies were associated with increased reports of depressive symptoms. Observations of mother-teen problem-solving revealed that depressed teens engaged in interactions characterized by high levels of maternal dominance and dysfunctional anger. Maternal reports of teens' negative life events contributed additional variance to a regression model for depressive symptoms. Gender differences further specified pathways to depressive symptoms. Females reported more depressive symptoms, while depressed males engaged in problem-solving interactions characterized by high levels of dysfunctional anger. Competency-based treatments of depressive symptoms designed to improve parent-teen communication and meta-monitoring skills are discussed.

Forty adults with serious psychopathological disorders were administered the Adult Attachment Interview (AAI). Attachment strategies were assessed using Kobak's (1989) Q-sort method, which yields scores along the two dimensions of security/anxiety and repression/preoccupation. Expressed emotion was rated for a close family member based on the Five-Minute Speech Sample (FMSS). In addition, premorbid competence was assessed, and subjects completed a psychiatric symptom inventory. Subjects who used more extreme secondary attachment strategies (whether preoccupied or repressing) were more likely to have family members who were overinvolving. Symptom inventories revealed greater symptom reporting by individuals who relied more on preoccupied than on repressing strategies and by individuals who were in high expressed emotion families. Higher levels of premorbid competence were associated with more secure attachment strategies. These findings are consistent with a transactional model of relationships whereby the attachment strategies of these adults affect the involvement of their family members, which in turn further perpetuate behaviors associated with their attachment strategies.

The relation among family relationships was explored for (a) couple pairings and (b) parent-child dyads. In S3 maltreating and adequate families, mothers' and male partners' quality of attachment (drawn from interviews) was compared with each other and with child quality of attachment (drawn from the Strange Situation). The notion of internal representational models, drawn from attachment theory, provided the basis for seeking continuity, that is, matches and meshes between partners and coherence from parent to child. Both concordance and discordance were found; in particular, a hypothesis of meshed adult relationships and parent-child transformations was supported. The conditions leading to meshed partnerships and transformed parent-to-child patterns require further study. In addition, the results suggest the need for greater incorporation of regulation of affect in assessment procedures, greater theory development, and precision in validation and application of assessment procedures. Finally, this study highlights the advantages of using atypical samples to explore aspects of normative development that are obscured in normative samples.

Attachment theory and family systems perspectives were used to explore the intergenerational origins of an overinvolved, symbiotic parenting pattern that hampers children's emerging autonomy. Consistent with principles of developmental psychopathology, the same underlying relationship disturbance was expected to manifest differently at different times in the developmental process. Specifically, links among maternal intrusiveness during infancy, overprotection during childhood, and mother-child role-reversal during young adulthood were explored. Forty-nine maternal grandmothers, mothers, and firstborn infants were visited in the mothers' homes when the infants were 6 and 9 months old. Grandmothers' memories of overprotection were related to observational ratings of high boundary dissolution and low positive affectivity with the mothers. No significant relationships were found between mothers' past and current relationships with the grandmothers. Yet, mothers' memories of overprotection during childhood or their current participation in unaffectionate, enmeshed, and role-reversed relationships with their mothers forecast their intrusive care with their own infants. Moreover, mothers' memories of overprotection related to their beliefs that others cannot be trusted and these beliefs were related, in turn, to observations of their intrusive care at 9 months. The implications of understanding the origins of maternal intrusiveness for developing prevention programs are discussed.