The secure working model classification of adult attachment, as derived from Main and Goldwyn's (in press) Adult Attachment Interview scoring system, was considered in terms of earned-security and continuous-security. Earned-security was a classification given to adults who described difficult, early relationships with parents, but who also had current secure working models as indicated by high coherency scores; continuous-security referred to a classification in which individuals described secure early attachment relationship with parents and current secure working models. Working models of attachment were classified as earned-secure, continuous-secure, or insecure in a sample of 40 parents of preschool children. Comparisons among the classifications were conducted on a measure of depressive symptoms and two sets of ratings of observed parenting styles. Adults with earned-secure classifications had comparable depressive symptomatology to insecures, with 30% of the insecures, 40% of the earned-secures, and only 10% of the continuous-secures having scores exceeding the clinical cut-off. The rate of depressive symptomatology in the earned-secure group suggests that reconstructions of past difficulties may remain emotional liabilities despite a current secure working model. With regard to parenting styles with their preschoolers, the behavior of earned-secure parents was comparable to that of the continuous-secures. This refinement in conceptualizing secure working models suggests ways for understanding variation in pathways to competent parenting as well as a possible perspective on how adults' adverse early experiences may continue to place them and their children at risk. Attachment theory as proposed by Bowlby The defining qualities of the working model (1973,1980, 1988) has offered a compelling are based on at least two judgments: (a) framework for considering continuity and whether or not the attachment figure is asmalleability in individual development and sumed to be a person who in general will resocial relationships. From repeated experi-spond to requests for support and protecences with a caretaker, a child constructs an tion, and (b) whether or not the child judges internal working model or representation of his-or herself as a person whom anyone, self, and of the self in relation to others, particularly the attachment figure, is likely

Twenty-seven mothers and 27 fathers were given the Adult Attachment Interview (M. Main & R. Goldwyn, in press) when their children were 3.5 years old. Continuous ratings of narrative coherence, probable experience quality (parents perceived as loving), and state of mind (current anger at parents) were entered as latent variables in partial least squares structural equation models that included observational measures of marital quality and parenting style. Models that include fathers' attachment histories predicted more variance in kindergarten teachers' descriptions of children's externalizing behavior, whereas models that include mothers' attachment histories predicted more variance in children's internalizing behavior. Marital data added predictive power to the equations. Discussion is focused on the importance of integrating attachment and family systems approaches, and of parents' gender and marital quality, in understanding specific links between parents' attachment histories and their young children's externalizing and internalizing behaviors.

Previous research has documented connections between adults' working models of childhood attachment relationships and the quality of parent-child relationships, but less attention has been devoted to examining such links for intimate adult relationships. Twenty-seven married couples were given George, Kaplan, and Main's Adult Attachment Interview and each person was rated as either secure or insecure with respect to attachment. Self-report measures of satisfaction with couple communication and marital relations and laboratory observations of couple interactions were collected. Results showed that self-reported marital satisfaction was not related to adult attachment classifications for either husbands or wives. However, observational ratings of couple interaction yielded differences for husbands. As compared to husbands classified as insecure, secure husbands were likely to be in better-functioning couples who engaged in more positive and fewer conflictual behaviors. In addition, couples' joint attachment classifications were related to observed couple behavior. Insecure-secure and secure-secure dyads did not differ, but both groups showed less conflict and were rated as better functioning than were insecure-insecure dyads. These findings suggest that a secure partner may buffer the negative effects of insecure attachment on the marital relationship.

This study addresses the question of whether or not parents' working models of childhood attachments constitute a risk factor for difficulties in current parent-child relations. In a sample of 27 families and their preschool-aged children, mother-child and father-child dyads were observed in separate laboratory play sessions from which ratings of parents' and children's behavior were collected. Working models of attachment were assessed using the Adult Attachment Interview (George, Kaplan, & Main, 1984). Results showed that parents classified as insecure were less warm and provided less structure in interactions with their children than did parents classified as secure. Children of insecure parents were less warm toward their parents than were children of secure parents. Analyses of parents' joint attachment classification showed that insecure women married to insecure men were less warm and provided less structure with their children than did mothers in either the insecure-secure or secure-secure dyads. These findings suggest that, in two-parent families, an insecure working model may be a risk factor for less competent parenting but that the risk is more pronounced when both parents have insecure working models of attachment.

89 children and their mothers participated in a study examining the association between attachment and peer social competence. During the summer following kindergarten, quality of attachment was assessed from reunion episodes following a 1-hour separation. In the fall, measures of sociometric status, peer behavior nominations, and peer liking ratings were collected. Teachers completed liking ratings and ratings of behavior problems and competence. Consistent with longitudinal studies of infant attachment and peer relations, insecurely attached boys were less well liked by peers and teachers, were perceived as more aggressive by classmates, and were rated by teachers as less competent and as having more behavior problems than were their secure counterparts. No such associations emerged for girls. Possible explanations for unanticipated differences in the pattern of results for boys and girls are discussed.