Promotive and protective factors for young children in three large representative samples: Comparing individual and cumulative models

• In Event: Multisystem Resilience Factors: Promoting and Protecting Children’s Mental Health in Multiple Risk and Cultural Contexts

Abstract

Early childhood is a critical developmental period that lays the foundation for future learning, behavior, and health. Drawing on theories of risk and resilience, we compared methodologies evaluating how positive factors may be either promotive (equally benefitting all individuals regardless of risk) or protective (buffer the effects of risk) for decreased problem behaviors in early childhood.

Methods: Data were combined from three national datasets (N=15,950): NICHD Study of Early Child Care and Youth Development (SECCYD), the Early Childhood Longitudinal Study- Birth Cohort (ECLS-B), and the Fragile Families and Child Wellbeing Study (FFCWS). Eight potentially positive factors were identified across datasets: (i) child social competence (ii) attachment security, (iii) parental sensitivity, (iv) parental efficacy, (v) good parent mental health, (vi) rated home quality, (vii) public service usage, and (viii) material security (food, housing, material). Five widely used demographic risk factors were summed into a cumulative risk index, and both models included covariates for child sex, race/ethnicity, and age. This paper compared two commonly used models: an individual effects model assessing the unique influence of positive factors and the potential buffering effect of the interaction of each variable and risk; and a cumulative model using summed scores for both positive factors and risk scores, along with their interaction.

Results: Both models demonstrated beneficial effects of positive factors on decreased problem behaviors. In the individual effects model, several positive factors had negative main effects on problem behaviors (promotive effect). When interactions were tested one at a time, both attachment security (b = -.38, p < .05) and home quality (b = -.34, p < .05) had a significant interaction with the risk index. When entered simultaneously in a single model, attachment security x risk remained significant (b = -.36, p < .05), and home quality x risk trended towards significance (b = -.34, p = .06), accounting for 17.4% of the variance in total problem behaviors (Table 1).

When public service use was included in the protective index, the risk and protective index variables had independent main effects on problem behavior (b = 1.41, p < .001 for risk index and b = -9.51, p < .001 for protective index). The risk and protective indices did not interact in their effects. When public service use was not included in the protective index, the indices interacted in their effects on problem behaviors (b = -.84, p < 05). Figure 1 demonstrates that a higher risk index was more strongly associated with problem behaviors among youth who scored low on the protection index (b = 1.35, p < .001), compared to youth who scored high on the protection index (b = 1.00, p < .001), accounting for 12% of the variance in problem behaviors. The implications for both models in methodology, policy, and practice are further discussed. Intervention efforts inclusive of supporting both promotive and protective factors may be beneficial.

Authors

• Andrew (Drew) B. McGee, University of Denver
  Presenting Author

• Julia Dmitrieva, University of Denver

• Angela J. Narayan, University of Denver

• Sarah E Watamura, University of Denver