lation are provided and only the availability of stable caregivers is lacking, development is substantially less delayed (Hodges and Tizard, 1989a, 1989b; Tizard and Hodges, 1978; Tizard and Joseph, 1970; Tizard and Rees, 1974). This suggests that stable relationships, as distinct from social-cognitive stimulation, are not required to ensure adequate physical, sensorimotor, cognitive, and language development. Children who have lacked stable and consistent caregiving, however, are not free of problems. Using parent and teacher reports, these children have been found in several studies to show impairments in regulatory aspects of thinking that involve concentration, attention regulation, and inhibitory control, generalizing problem solutions, and excessive concreteness of thought (Hodges and Tizard, 1989b; Tizard and Hodges, 1978, as reviewed in Gunnar, in press).
Importantly, removing children from institutions and placing them in stable families with adequate resources can produce remarkable catch-up growth on developmental milestones and in general cognitive (i.e., measures of IQ) and language development (Ames, 1997; Rutter, 1972, 1981b; Rutter and the English and Romanian Adoptees [ERA] Study Team, 1998). Even children delayed a year or more in behavioral and physical development can achieve normal levels of functioning once they are given the opportunity to live with a loving family.
Nonetheless, a persistent minority of institutionalized children across all studies and samples fails to show this dramatic recovery. They continue to exhibit multiple, debilitating problems in cognition and behavior years after entering their new families. Multiple, often unknown factors are likely to constrain developmental outcomes for this persistent minority. Case by case, these factors may include varying mixtures of genetic, prenatal, and postnatal conditions. If institutional rearing is involved, the continuing deficits found in some children should show a consistent dose-response relationship with the duration of privation experienced. Duration of orphanage exposure is highly confounded with illness, maltreatment, repeated changes in caregiving, and so on, making it exceedingly difficult to isolate duration as the causal factor. However, recent findings based on the Romanian children offer a reasonable test of this hypothesis (O'Connor et al., 2000; Rutter and the English and Romanian Adoptees [ERA] Study Team, 1998). In fact, a strong dose-response relationship was found (with a 24-point IQ difference between the extremes), in which children who spent more time in the institutions showed more persistently poor outcomes. The study controlled for a series of alternative explanations, including the possibilities that later-adopted children might have had greater impairments at birth and that the differing outcomes may have been due to the children's degree of malnutrition, which adds to the credibility of the conclusion.
The outcomes of institutionalized children may also be constrained by