On individual differences in strange situation behavior: Categorical and continuous measurement systems in a cross-cultural data set

Recent discussions of attachment research methodology have questioned the appropriateness of measuring individual differences in Strange Situation behavior in terms of the avoidant, secure, and resistant categories. Continuous, factor-analytically derived latent variables have been proposed as an alternative. The validity of the traditional categorical measurement system was investigated through a cluster analysis of 216 11- to 13-month-old American, Swedish, and Israeli infants' social interactive behaviors in the reunion episodes of the Strange Situation. These analyses indicated that differences within the secure group based on variation in proximity-seeking and contact-maintaining behavior are more prominent than differences between the secure infants and the insecure infants. Moreover, similarity among infants was based on the infants' cultures as well as their membership in the avoidant, secure, and resistant groups. The alternative of continuous, latent-variable measurements was studied using confirmatory factor analyses on the same data. These analyses demonstrated that there are at least two dimensions in the infants' reunion behavior, but the structure of these dimensions varied across cultures. The results of both analyses suggest that there are substantial cultural differences in the patterning of individual differences in Strange Situation behavior.