Following the birth of an infant it was thought that differences in maternal self-esteem (MSE) could be determined from individual differences in infant risk status (IRS), infant behavior (IB), and social support (SS). Previous research has not delineated the relative importance of a mother's perception of infant health (POIH), the assessment of temperament, and intimate SS compared to actual medical risk status, neonatal behavior, or total SS in predicting MSE in the neonatal period. A related purpose was to examine the complex nature of MSE within the emergent process of social interaction using both symbolic interaction and the transactional model of development. The central question proposed that the perceptual variables would precede the objective measures of the same variables in all categories.

The total sample (TS) included 77 mother-infant dyads. Mothers were healthy, $>$16 years, middle-lower socioeconomic status. Infants included 36 healthy, neurologically, normal, full-term (FT) infants, and 41 preterm (PT) infants, appropriate for gestational age, with a range of problems reflecting early medical course.

Medical risk was assessed by the Hobel Scale at discharge. POIH was measured via questionnaire at recruitment. Neonatal Behavioral Assessment Scale (NBAS) 40 week data, measured IB. At 44 weeks corrected gestational age infant temperament, via the Bates (ICQ), SS from a partner/significant other, and total available SS assessed from the Norbeck Social Support Questionnaire (NSSQ) predicted MSE measured by the Maternal Self-report Inventory.

Six of the seven hypotheses generated from the central question were supported. Significant correlations were found in the expected direction for 4 variables in the IRS and IB categories. POIH was independent of medical risk status in its relationship to MSE. Partner was perceived to be the most important SS to a mother. SS did not have a relationship with MSE. Support for the central question was also found. POIH, ICQ-fussy, and NBAS autonomic cluster accounted for 48%, $p$ $<$.0001, of MSE for the TS. Hierarchical regression supported the theoretical order of the categories and variables.

MSE, based on gestational age, indicated POIH and ICQ-fussy accounted for 25%, $p$ $<$.01, of MSE in FT mothers. NBAS range of state, ICQ-fussy, and POIH accounted for 22%, $p$ $<$.02, of MSE in PT mothers.