Method: Design used a secondary data analysis for information collected during a longitudinal service-project. Families were recruited from local prenatal obstetric-clinics. Current study used 198 non-white (53-percent Latino, 47-percent African American) families with complete data, who qualified to receive project services (focus-child had medically normal birth; mother had GED or less education but able to speak basic English/Spanish; family lacked health insurance), and child's age ranged 16-18 months. Trained (bilingual) interviewers collected data during structured interviews. The study measured maternal depression by EPDS (Cox, Holden & Sagovsky, 1987); parenting knowledge by KIDI (MacPhee, 1981); constructed a Likert-scaled variable of risky parenting practices from dichotomous-questions concerning parents' use of corporal punishment, home safety, and recommended well-health visits (AAP 2002); stimulation activities by direct-literacy questions (Newman & Dickinson, 2002) and indirect-enrichment-literacy questions (AAP, 2002); and child language by MacArthur-CDI (Fenson et al., 2000). SPSS-AMOS 16 analyzed proposed SEM path-models. Maximum-likelihood calculated path coefficients. SEM models were tested for overall fit-measures of the chi-square (p>.05); NFI, CFI, and RMSEA.
Results: Findings provided support for a mediated path-model with a good fit: Chi-square(16, N = 198) = 16.72, p = .401; NFI = .97, CFI = .99, RMSEA = .01. Final model demonstrates that children's language is impacted by maternal depression indirectly. Maternal depression significantly affected parental knowledge of infant development and risky parental practices for well-health, home safety, and corporal discipline. Risky parental practices are affected by maternal depression and knowledge directly and indirectly. The impact of maternal depression on parental practices is also mediated by parental knowledge. Modeled paths explained 24% of variance in risky parental practices. Children's involvement in stimulation activities can be predicted directly and indirectly from parental knowledge, risky parental practices, and maternal depression. The impact of maternal depression on stimulation activities is also mediated by parental knowledge and risky practices. Paths explained 35% of variance in children's involvement in stimulation activities. Children's involvement in stimulation activities directly affected their emergent vocabulary development while mediating the impact of parental knowledge, risky parental practices, and maternal depression. Paths explained 42% of variance in children's emergent vocabulary.
Implications: Results suggest although social work practitioners need to promote adequate parental knowledge and healthy parental practices, they also need to strive to reduce maternal depression while assuring healthy literacy-oriented stimulation of children. Programs providing early educational and family support need to train staff in recognition of maternal depression while promoting parental knowledge and practices toward healthy child development.