Multiple placement is the characteristic most associated with negative outcomes for youths in out-of-home care, two to five years following emancipation [Cook in press]. Of 14 variables associated with outcomes, multiple placement was negatively associated with 12 and neutral for two. Negative outcomes include unemployment, school dropout, relationship troubles, and teen parenthood. A history of placement in out-of-home care is a risk factor for becoming homeless [Piliavin et al. 1992]. Multiple placements and emancipation from group care, rather than family situations, are associated with incarceration in young adulthood [Fanshel 1990; Festinger 1983].
The issues involved in discharge planning for this vulnerable population necessitate understanding who these young people are, and how they experience attachment relationships and transitions. Emotionally disturbed adolescents in out-of-home care are most likely to have histories of placement disruption, particularly those adolescents with externalizing disorders [Pardeck 1983; Proch & Tabor 1987]. These disorders include attention deficit hyperactivity disorder, oppositional-defiant disorder, and conduct disorder [American Psychiatric Association 1987]. A lack of stable parent figures most severely affects social development. As attachments become more and more tenuous, children become less selective regarding relationships, so that they are likely to drift into harmful relationships [Rutter 1993] or to repeat the pattern that has already been established of drifting through relationships.
Bowlby  suggests that as the child grows older the pattern of attachment--including the personality that develops from it--tends to establish communication patterns that resist change. Thus, the child tries to impose previously experienced patterns on new relationships. Patterns of attachment tend to continue so that securely attached infants of 12 months become six-year-olds who have relaxed and friendly relationships with their parents, as well as with other adults and peers. Children who develop insecure attachments may be classified as anxious-resistant or anxious-avoidant [Ainsworth 1978]. The latter is analogous to the position Bowlby  ties to literal or figurative abandonment and so, provides the conceptual focus for our analysis.
The construct of attachment shows both continuity and change over development, but can be predicted by infant attachment status [Sroufe & Egeland 1991; Sroufe et al. 1990]. Anxiously attached children at 42 months are "less ego-resilient, less independent, less compliant, less empathic, less socially competent, lower in self-esteem" than securely attached children [Waters et al. 1985: 112]. Anxious-avoidant children are described by their preschool teachers as hostile and socially isolated from peers [Sroufe & Egeland 1991]. In another study, securely attached preschool children asked to judge the intentions of children in cartoon stories were found to make realistic attributions, whereas anxious-avoidant children were more likely to attribute hostile or negative intent to the children [Elicker et al. 1991].
Renkin et al.  found "a clear tendency for avoidantly attached boys to be over-represented in the aggressive group." An association appears to exist between antisocial preschool behavior problems and a history of anxious attachment [Waters et al. 1985: 110]. Like the construct of attachment, secure readiness to learn "also appears to represent a dynamic balance between establishing safe, secure relationships with adults and feeling free enough to venture out to explore the world in a manner that is likely to promote maturation of cognitive competencies" [Aber et al. 1989: 594]. Thus, avoidant children are prone to school learning problems.
The interpersonal problems of behaviorally troubled children have been repeatedly documented in association with histories of anxious-avoidant attachment [Elicker et al. …