Ainsworth (1967; Ainsworth et al. 1978) and Bowlby (1969) identified four phases in the development of child-parent attachment. Over the first few months of development, children become increasingly active in maintaining the attachment relationship. By the time they have developed a clear-cut attachment towards the end of the first year, the infant has developed a pattern of relating to his primary caregivers based on his experiences with them. When they reach the fourth phase, around age 3 ½ or 4, the child-parent relationship has become a “goal-directed partnership” where true conflict and cooperation is possible. Descriptions of the phases of attachment development are summarized in Table 2.2.

In addition to looking for indicators of when a child became attached to a caregiver, Ainsworth used observations from her field studies to describe the quality of the attachment relationship. Ainsworth (1967) initially referred to this as the strength or security of the attachment relationship but in later writings exclusively used the term security (Ainsworth et al. 1978); “The obvious first impulse was to try to assess strength of attachment, but this ran up against a brick wall when one realized that this could not be achieved by the mere assessment of the strength or intensity of attachment behavior, for this is situational, and furthermore it is those who are anxiously attached who tend to have the strongest attachment behavior in the natural environment” (Ainsworth 1988).

Based on her extensive naturalistic observations of infant-mother interactions, Ainsworth et al. (1978) developed a standardized observational assessment known as the Strange Situation Procedure (SSP) to examine individual differences in the quality of infant-mother attachment. This procedure involves a series of situations designed to assess the balance between exploring the environment and seeking proximity to the attachment figure. The parent and child enter a novel playroom with attractive toys (designed to activate exploratory behavior). The baby is then confronted with a series of increasingly stressful situations designed to activate attachment behavior: the entrance of a stranger who first talks to the mother and then initiates interaction with the baby, an initial brief separation where the mother leaves the child with the stranger, and a second brief separation where the child is left alone in the room. Each situation lasts 3 min. However, separations from the caregiver are curtailed early if the infant becomes distressed.

Based on observations of different aspects of the infant’s or child’s behavior in the SSP (e.g. proximity-seeking, contact-maintaining, avoidance, and resistance), the dyad is given a classification that summarizes the quality or pattern of attachment. In evaluating the security of the attachment relationship, the child’s response to reunions with the mother following separations is especially important. Despite the complexity of the coding system, interrater reliability is good when the SSP is coded by well-trained observers (e.g. 80–90 % agreement; kappa = 0.69 to 0.72) (Cassidy et al. 2011; NICHD Early Child Care Research Network 1997). Information about training in coding the SSP is available at the Attachment Training web site: http://​attachment-training.​com/​at/​. The SSP is considered the “gold standard” for assessing the quality of infant-parent attachment due to the extensive research supporting its relationship to in-home observations and longitudinal outcomes (Zeanah et al. 2011).

The three patterns of attachment originally identified by Ainsworth et al. (1978) are now referred to as organized patterns of attachment. When Ainsworth first described these different patterns of attachment, she assigned a letter to each type, a shorthand that has stuck over the years. Securely attached infants and young children (B) exhibit a balance between proximity-seeking and exploration. They directly communicate distress in situations that provoke uncertainty or fear, seek proximity to their mother when distressed, are soothed by their mother, and return to exploration. In the other two organized patterns of attachment, the baby or young child is primarily focused on seeking and maintaining proximity to the primary caregiver (insecure-ambivalent/resistant) (C) or is primarily focused on exploring the environment (insecure-avoidant) (A) rather than exhibiting a balance between proximity-seeking and exploration. In the first four samples where Ainsworth used this classification system (N = 106), 66 % of dyads exhibited a secure attachment (B), 22 % exhibited an avoidant attachment (A) , and 12 % exhibited an ambivalent/resistant attachment (C) (Ainsworth et al. 1978) .

The classification of disorganized attachment (D) was developed by Main and Solomon (1990) after reviewing videotapes of dyads that were difficult to “fit” into Ainsworth’s organized attachment classifications. The infants in these dyads exhibit a variety of conflict behaviors in stressful situations when they are in the presence of their caregiver. These conflict behaviors are not consistent with the organized patterns identified by Ainsworth and came to be understood as a breakdown in the infant’s ability to effectively use the mother for emotional regulation under stressful circumstances (Main and Solomon 1990; van Ijzendoorn et al. 1999). Breakdowns sufficient to classify a dyad as disorganized can range from brief interruptions in an otherwise organized pattern of attachment to global disorganization.

The majority of infant-mother dyads classified as disorganized are given a secondary, best-fitting Ainsworth et al. (1978) classification. For example, an infant classified as disorganized/secure² (D/B) might exhibit an overall pattern of going to their parent for comfort when distressed but exhibit conflict behaviors in the context of comfort-seeking (e.g. briefly freezing or engaging in stereotypies such as hair twisting or rocking on the approach to the parent or turning around and backing toward the parent for comfort). In the relatively rare circumstance where the dyad exhibits global disorganization and a secondary classification cannot be determined, the dyad is classified as disorganized/cannot classify.

The SSP and coding system has been modified for use with young children (ages 2 through 6) (Cassidy et al. 1992) and 6-year-olds (Main and Cassidy 1985). There is a growing body of literature on the modified SSP for preschool-aged children, making it the best-validated assessment of attachment security for young children (Greenberg et al. 1991; Moss et al. 2004; Speltz et al. 1990, 1995, 1999). Like the infant SSP, the preschool adaptation uses brief separations from the caregiver. In early studies, the infant procedure was followed but separations were lengthened. In later studies, there continue to be longer separations than used for infants (5 min vs. 3 min) but a stranger is not used. The modified SSP for pre-schoolers used in current studies is as follows: The parent and child enter a playroom for 5 min. The parent then leaves for 5 min, returns for 5 min, leaves a second time for 5 min, and returns a second time for 5 min (Moss et al. 2004, 2011). The procedure for 6-year-olds uses a 1 h separation where the child is with a stranger and only one reunion (Main and Cassidy 1985).

When evaluating quality of attachment in infants and young children, it is important to recognize the major developmental shifts that occur between infancy and early childhood. For example, while securely attached infants often cry during the 3 min SSP separations, securely attached pre-schoolers rarely cry during 5 min separations.

Striking developmental shifts from infancy to school-age are seen in longitudinal studies of infants with a disorganized attachment relationship (Hesse and Main 2000). Many of the school-aged children who display the conflict behaviors and anxiety indicative of disorganized attachment as infants display a controlling pattern of interacting with their parent as 6-year-olds. It is as though they have resolved their conflict and anxiety by taking charge of the relationship. Details of the patterns of attachment in infants and pre-schoolers are presented in Table 2.3.

The parenting behavior identified in Ainsworth’s research as central to the development of secure attachment is sensitive responsiveness, i.e. the parent’s prompt, consistent, and appropriate response to the baby’s attachment signals (Ainsworth et al. 1978). As shown in Table 2.1, attachment signals include both affectively positive signals (e.g. smiling, reaching, following) and affectively negative signals (e.g. crying). Subsequent research suggests sensitive responsiveness to infant distress is especially important in the development of a secure infant-parent attachment (Del Carmen et al. 1993; van den Boom 1988, 1989, 1994). However, it is important to note the distinction between sensitive responsiveness and responsiveness. Sensitive responsiveness involves both knowing when to respond to a baby and when to let the baby utilize their emerging capacities for self-regulation (Beebe et al. 2010; van IJzendoorn and Hubbard 2000). Research indicates a curvilinear relationship between responsiveness and secure attachment suggesting sensitive responsiveness involves the “just right” amount of responsiveness – neither too much nor too little (Beebe et al. 2010).

The parenting behavior with the strongest empirical support for its association with attachment is sensitive responsiveness. However, this parenting behavior explains only a portion of the variance in security of attachment. Table 2.4 summarizes some of the other parenting behaviors associated with the development of secure attachment .

Insecure attachment is clearly a risk factor for disruptive behavior; a meta-analysis of 69 studies examining the association of attachment and externalizing behavior problems found a statistically significant association (d = 0 .31) (Fearon et al. 2010). Disorganized attachment had a stronger association with externalizing behavior problems (d = 0 .34) than avoidance (d = 0 .12) or resistance (d = 0 .11).

Numerous interventions targeting infants/young children and their parents cite attachment theory as a theoretical foundation and indicate their intervention improves attachment (Bakermans-Kranenburg et al. 2003, 2005; Bernard et al. 2012; Cassidy et al. 2011; Chaffin et al. 2011; Cohen et al. 1999; Dozier et al. 2002, 2007; Eyberg 2005; Hoffman et al. 2006; Ijzendoorn 1995; Moss et al. 2011; Sanders 2010; van den Boom 1988, 1989, 1994). This review focuses on interventions with outcome studies utilizing research-based observational assessments of attachment security (SSP or Modified SSP for Preschoolers).

Watch, Wait, and Wonder (WWW), Infant-Parent/Child-Parent Psychotherapy (CPP), and Circle of Security (COS) are described below and outcome studies of these interventions are summarized in Table 2.5. Watch, Wait, and Wonder (WWW) is the only attachment-based intervention demonstrated to ameliorate insecure attachment among infants and toddlers referred for mental health concerns (Cohen et al. 1999). WWW is a dyadic, child-led approach based on attachment theory and object relations theory (specifically, Winnicott and Bion’s concepts of the holding environment and projective identification). During the first half of the session (20–30 min), the mother is encouraged to get down on the floor with her infant or child and follow his lead. The mother is told if her child initiates an interaction with her she should respond, keeping her child’s agenda in mind and not introducing her own agenda. She is told that if she is not sure what to do, she should remember “Watch, Wait, and Wonder.” In the second half of the session (20–30 min), the therapist asks the mother about her observations of the child, her experience of doing child-led play, and any difficulties she had doing child-led play.

Infant-Parent Psychotherapy/Child-Parent Psychotherapy (CPP) is an intervention developed by Selma Fraiberg (CPP became the inclusive title when Infant-Parent Psychotherapy was extended to toddlers and preschool-aged children) (Fraiberg 1980; Fraiberg et al. 1975; Lieberman and Van Horn 2008). Fraiberg, who was trained as a social worker and psychoanalyst, developed CPP as a home visiting model focused on improving infant-parent interactions. The premise of this approach is disturbances in child-parent relationships result from the parent’s psychological conflicts being expressed through the parent’s attitudes and behaviors toward the infant and the best way to address these disturbances is working directly with the child-parent dyad. The primary modality is joint child-parent sessions where spontaneous child-parent interactions are used to support the child’s positive development and address parental conflicts that may be a barrier to healthy development. Some of the strategies therapists use include “speaking for the baby” to help the parent recognize the meaning of the child’s behavior, developmental guidance, crisis intervention, and helping the parent recognize how exposure to traumatic events has impacted their interactions. A core theoretical tenet of CPP is that when the parent is able to remember the affect associated with a painful event they are able to become protective of their child rather than acting out past traumas in their interactions with their child. In the outcome studies reported in Table 2.5, CPP lasted 5–12 months (number of sessions ranged from 15–45). As shown in Table 2.5, CPP leads to significant increases in security of attachment in dyads with maltreating parents (Cicchetti et al. 2006) and mothers with Postpartum Major Depression (Toth et al. 2006). However, in the only study examining changes in attachment security in a clinically-referred population, families receiving CPP exhibited a decrease in secure attachment following treatment (Cohen et al. 1999).