Instability of infant–parent attachment security.

According to attachment theory, the infant during the second half of the first year of life develops an internal representation of the attachment relationship. Once formed, this representation is relatively stable (although still subject to modification based on subsequent experience) and guides the infant's behavior in new situations. The Ainsworth and Wittig (1969) Strange Situation assesses infant attachment and, with some exceptions (e.g., in high-risk samples or when child care arrangements change markedly), Strange Situation classifications are presumed to be relatively stable during the second year. Most American research relevant to this hypothesis was conducted over a decade ago and showed that for a substantial minority of infants, instability characterized the development of attachment security during the second year. Multistudy data compiled by Thompson (in press) reveal that among working- and middle-class samples in the United States not considered at-risk for contextual (e.g., poverty, maltreatment) or biological (e.g., prematurity) reasons, stability in infant-mother attachment security between 12 and 1823 months ranges from 53% (Thompson, Lamb, & Estes, 1982; n = 43) to 96% (Waters, 1978; n = 50). In fact, when rates of stability of ABC classifications from five independent investigations of nonrisk American families are weighted by sample size (Frodi, Grolnick, & Bridges, 1985; Main & Weston, 1981; Owen, Easterbrooks, Chase-Lansdale, & Goldberg, 1984; Thompson et al., 1982; Waters, 1978), a 75% rate of stability emerges on a total sample of 205 infant-mother dyads. The comparable figure for infant-father attachment classifications based on data from two of these investigations (Main & Weston,