125P
A Latent Class Analysis of Resilience Factors Among Children At Risk for Maltreatment
Research on resilience and adaptation among maltreated children over the past two decades has shed light on the positive outcomes among these children (e.g., Jaffee et al., 2007). Existing studies on children’s resilience are based on numeric cutoffs of composite scores comprised of multiple indicators (Flores et al., 2006), which has multiple methodological drawbacks. These include largely ignoring heterogeneity inherent in response to different factors leading to resilient functioning, failing to discern differences within various factors, and simplifying the complexity of the resilience concept. Given that children’s adaptation reflects in various areas, this study intends to provide a more refined characterization of resilience factors using latent class analysis.
Methods:
Data were drawn from the Longitudinal Studies of Child Abuse and Neglect (LONGSCAN), a cohort study following nearly 1,300 children at varying levels of risk or exposure to maltreatment across the country. Data for this study were taken from a survey examining children’s resilience factors when the children were 12 years old in four broad domains, including existence of supportive adults, religious institution involvement, prosocial extracurricular activities, and history of leadership (N=813). Based on children’s responses to these 17 dichotomous items (i.e., yes or no), latent class modeling was used in Mplus 6.1 to reveal the number of distinct latent classes and the prevalence of each class. Multinomial logistic regression was then estimated to examine relationships between socio-demographic factors and latent class memberships.
Results:
Findings indicate that the 3-class model was the best fitting solution (entropy=0.83), representing different patterns of these children’s resilience factors. The first class represented nearly half (47.4%) of all children, and was characterized by high scores in relationships with supportive adults and moderate scores in involvement in religious and extracurricular activities. The second class was characterized by high values across all the resilience factors, representing 44% of those children. Finally, the third class was characterized by low values in relationships with supportive adults and moderate scores in involvement in religious and extracurricular activities, representing 8.6% of the total sample. Non-white children were less likely to have high values in supportive relationships (Class 1: OR=0.49, p<0.05) and high values across all resilience factors (Class 2: OR=0.37, p<0.01) compared to white children. Receiving social welfare services (e.g., TANF) was associated with a higher likelihood of having more supportive adults and moderate extracurricular activities (Class 1: OR=1.13, p<0.05).
Conclusions and Implications:
Results suggest that at-risk children’s resilience factors can be categorized in different classes, confirming the heterogeneity in children’s overall resilient functioning and adaptation. The findings support the need to create group memberships, particularly when using dichotomous measurements. These different classes of children could benefit from specially tailored services that acknowledge variations in different resilient aspects. For the largest group who have supportive adults but low extracurricular activities, social workers can strengthened their affiliations or belonging to social groups. Minority children at risk for maltreatment tend to lack resilience factors and future studies are needed to further explore the reasons and effective interventions to promote their resilience.