Waters reports on the examination of 50 infants from the Minnessota Birth Announcement study. The Strange Situation Procedure was used at 12 and 18 months to investigate the longitudinal stability of infant-mother attachment across this period.   At 12 months, 10 infants were considered avoidant, 30 secure, and 10 ambivalent. At 18 months, 48 of the 50 infants (96%) were assigned to the same attachment style as had been recorded at 12 months. Both of the infants who changed attachment style moved into the secure group, one from the avoidant group and one from the ambivalent.

Analysis of discrete behavioural traits did not show temporal stability so this study lent support to the conceptualisation of attachment as an organisational construct, as previously argued by the author (Sroufe & Waters, 1977).

Contrast this with the results found by Vaughn, Egeland, Sroufe & Waters (1979) using subjects from poor, unstable families, and Egeland & Sroufe (1981) who used a sample of infants with neglectful mothers.