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, 1 6 female) were observed in multiple naturalistic and single free-play interactions at 1-, 4- and 9-months infant age: setting and coding pro cedures were held constant. Infant-mother attachment quality was asses sed at 1 year. Analyses examined similarities among and differences be tween data derived from the two interactional contexts, and the relati ve utility of each in explaining individual differences in attachment quality. Data from individual naturalistic observations were not as re liable as data consolidated across naturalistic observations. Addition ally, maternal sensitivity, as measured by repeated naturalistic obser vations, was far superior to sensitivity measured in the free-play con text in accounting for individual differences in attachment quality - securely attached infants experienced higher levels of maternal sensit ivity than did insecurely attached infants. Conversely, it was difficu lt to distinguish between naturalistic and free-play measures of mater nal rejection in terms of their ability to account for individual diff erences in attach ment - at 1 month, insecure-resistant infants experi enced the highest levels of maternal rejection.