Over the past 25 years, changes in society at large--increased acceptance of single-parent families, greater availability of contraception, the legalization of abortion--and changes in child welfare practice--increased abuse and neglect reporting and the implementation of permanency planning--have combined to change the shape of adoption practice. Of the approximately 20,000 children who are adopted via public agency auspices each year, more than half have special needs [Tatara 1988]. Key special-needs characteristics include older age at adoption (older than about age six, although ages vary by state), handicap, serious medical problem, emotional or behavioral problem, need for placement as a sibling group, and minority ethnicity.
SPECIAL-NEEDS ADOPTION RESEARCH
As special-needs adoption has increased, so also has the percentage of all adoptions ending in adoption disruption (the termination of adoption prior to legalization). A substantial body of research deals with disruption, while a somewhat smaller number of studies investigate the experiences of children and families in adoptions that remain intact. Kadushin and Martin's  summary of 11 studies of special-needs adoption found a range of disruption rates from 6% to 45%, with an overall rate of 11.3% (502 disruptions in 4,443 placements). The following predictors of increased risk of disruption were identified in a recent review [Rosenthal 1993: 81-82]:
older age of child at adoptive placement;
inadequate background information or unrealistic parental expectations;
rigidity in family functioning patterns, in particular the father's noninvolvement in parenting tasks;
low levels of support from relatives or friends; history of physical and particularly sexual abuse prior to adoption;
psychiatric hospitalization prior to adoption;
acting-out, externalized behavioral problems including sexual acting-out; and
adoptive placement with new adoptive parents rather than adoption by foster parents.
Findings from three studies of intact special-needs adoptive families converge to show that 75% or more of the parents were well satisfied with their adoption experiences [Kadushin 1970; Nelson 1985; Rosenthal & Groze 1992]. Particularly good outcomes are evident for children with developmental disabilities. For instance, in a recent British study, 95% of mothers who adopted children with developmental disabilities responded affirmatively when asked whether they would adopt if they could "do it over again" [Glidden 1990]. One national study found a disruption rate of only 3.3% for children with disabilities aged seven and younger at adoptive placement [Coyne & Brown 1985].
Longitudinal research on special-needs adoption is limited in breadth and scope. Important research includes Simon and Altstein's [1977, 1981, 1987] and Feigelman and Silverman's  studies of transracial and other adoptions; Shireman and Johnson's studies on single-parent and transracial adoptions [Shireman & Johnson 1976, 1985, 1986; Shireman 1988]; Westhues and Cohen's disruption study ; and Groze's work [in press] with a subset of the current study's sample. The recently implemented California Long-Range Adoption Study is beginning to follow a large sample of adoptees, some of whom have special needs [Barth 1991].
PURPOSE OF THE STUDY
The present longitudinal study examines changes in child and family functioning using a three-state mailed survey of adoptive families (most of whom adopted children with special-needs). The sample comprised families who responded to two administrations of the survey (N=302 usable responses). The mean age of study children was 10.2 years at the first survey administration (SD=3.0 years) and 13.5 years at the second administration (SD=3.1 years). Their mean age at initial entry to the home, often as foster children, was 4.7 years (SD=3.3 years). Surveys were completed by adoptive mothers or fathers. …