Studies that have concluded that human immunodeficiency virus (HIV) causes developmental delays have compared infected children's scores to instrument norms (Epstein et al., 1986; Ultmann et al., 1984), rather than matched control subjects. Early intervention programs have found similar deficits for children who were not HIV-infected but whose families were socially disenfranchised due to low SES and racial minority status. In the present study, cognitive (Bayley Mental Development Index), adaptive (Vineland Adaptive Behavior Composite), and physical (head circumference Z-score) measures were entered as outcomes in regression equations for low-SES samples of HIV-positive children (n = 9), children at risk for HIV who had seroreverted (n = 16), children too young to have conclusive HIV-antibody tests (n = 10), and children who had previously been admitted to a Neonatal Intensive Care unit (n = 19). The sample was predominantly African-American (84%) and half female. In addition to group membership, scores on the Parenting stress Index-Short Form (PSI-SF), Pediatric Review of Children's Environmental Support and stimulation (PROCESS), and a questionnaire about parent health were entered as predictors. Age differed across groups and was entered as a covariate. Group membership did not contribute significant variance to any regression. Age was inversely related to cognitive and adaptive development, accounting for 16% (p < .005) and 38% (p < .0001) in these respective outcomes. The parenting variables accounted for an additional 14% (p < .05) of the variance in cognitive development. No predictor was significant for head circumference. These results suggest that the delays observed in children with HIV may be attributed to environmental factors, rather than HIV as previous studies concluded. Children with HIV may thus be expected to benefit from early interventions as have their low SES peers. Generalization from these results remains tenuous, pending studies which replicate them while ameliorating methodological weaknesses of the present study. Future studies should match subject groups for age, control for experimenter bias, and utilize multi-site collaboration in order to obtain samples of sufficient size to test explanatory hypotheses about these developmental processes. / Ph. D.