Maternal attachment representations were assessed using the George, Kaplan, and Main (1985) Adult Attachment Interview (AAI), and emotional availability during observed mother-child interactions was assessed using the third edition of the Emotional Availability (EA) Scales (Biringen, Robinson, & Emde, 1998). This edition of EA included four parental scales and two child scales (Maternal Sensitivity, Structuring, Nonintrusiveness and Nonhostility; and Child Responsiveness and Child Involvement). Separate Hierarchical Multiple Regressions (HMRs) were computed to examine the prediction of the separate EA dimensions from demographic information, the AAI classification, and AAI scales. These analyses indicated that each of the EA dimensions (with the exception of maternal nonintrusiveness and nonhostility) was predicted by the AAI classification and/or AAI scales. Using three-step HMRs, the strongest prediction was for maternal sensitivity where 54% of the total variance in maternal sensitivity was explained by maternal education, AAI classification, and AAI 'state of mind' scales. Maternal nonhostility was predicted by maternal education and gender of the child, with lower-income mothers and mothers of girls demonstrating greater hostility.