Origins of Attachment: The Role of Context, Duration, Frequency of Observation, and Infant Age in Measuring Maternal Behavior

A total of 32 Caucasian women and their first-born infants (16 male, 16 female) were observed in multiple naturalistic and single free-play interactions at 1-, 4- and 9-months infant age: setting and coding procedures were held constant. Infant-mother attachment quality was assessed at 1 year. Analyses examined similarities among and differences between data derived from the two interactional contexts, and the relative utility of each in explaining individual differences in attachment quality. Data from individual naturalistic observations were not as reliable as data consolidated across naturalistic observations. Additionally, maternal sensitivity, as measured by repeated naturalistic observations, was far superior to sensitivity measured in the free-play context in accounting for individual differences in attachment quality-securely attached infants experienced higher levels of maternal sensitivity than did insecurely attached infants. Conversely, it was difficult to distinguish between naturalistic and free-play measures of maternal rejection in terms of their ability to account for individual differences in attachment-at 1 month, insecure-resistant infants experienced the highest levels of maternal rejection.