This study attempted to identify maternal variables within social class which were predictive of child status over time. Data on patterns of mother/child interactions were collected for the experimental and control groups of poverty mother/child dyads at the Frank Porter Graham Child Development Center and a group of general population high education and occupational status mother/child dyads. The measures employed included: (1) The Wechsler Adult Intelligence Scale or the Wechsler Intelligence Scale for Children and the High Risk Index administered prenatally; (2) a 25-minute laboratory observation of mother/child dyads in free play sessions conducted at 6 and 20 months; (3) the Caldwell Home Observation for Measurement of the Environment (HOME), administered in the subjects' own homes at 6 and 18 months; and (4) an adaptation of the Parental Attitudes Research Inventory (PARI), administered at 6 and 18 months. Results showed clear differences in the attributes of middle class and poverty mothers, with the general population mothers talking and interacting more with their infants and toddlers, scoring higher on all subscales of the Home Stimulation Inventory, and scoring as less authoritarian, more democratic, and more hostile and rejecting than the poverty mothers. It was concluded that within the poverty groups, day care intervention did alter the relationship between maternal variables and child status so that maternal variables were highly predictive of child status for the control group and less so for the experimental group. (JMB)