Purpose: Multi-stressed families often exhibit family dysfunction, but some families are resilient. They are able to maintain above average functioning despite experiencing increased stress. Measuring family resilience has been challenging. Past research includes a variety of definitions, studies of specific risk or protective factors rather than interactions between them, and individual outcomes or reports used to explain family processes. This study investigated relationships between a variety of risk factors, family protective factors, developmental outcomes, and demographic influences in order to explore the nature of family resilience.

Method: This study used data from a NIAAA funded (RO1 AA1164701A1), bi-national (United States and Canada), experimental research study of a family-based alcohol prevention intervention program. Families with at least one child between the ages of 9 and 12 and a parent with an alcohol problem were recruited into the study (N = 674). This descriptive study used pretest data to identify a typology of family protective factors and examine associations between that typology and risks, outside supports, developmental outcomes, and demographic characteristics. Risks were measured by a life events scale and current parent drinking. Protective factors included family functioning scales, measured by the Family Assessment Measure (FAM III), and presence of outside supports (professional services used, neighbor helpfulness, teacher caring). Parenting and child self esteem were assessed as family developmental outcomes. Family functioning and parenting measures combined parent and child responses to better represent family processes.

Results: Cluster analysis using five family functioning variables (beliefs, adaptability, cohesion, communication, and problem-solving) revealed three clusters of families. Families were either above average, average, or below average on all five variables. Relationships were found between cluster membership and teacher caring, χ² (8, n = 269) = 19.85, p < .05, four parenting measures, F (2, 267) = 19.7 to 33.5, p < .001, and family influence on child self esteem, F (2, 270) = 27.9, p <.001. There were no associations between the typology and risks, outside supports other than teacher caring, and most demographic characteristics (number of adults/children in home, child gender, SES, and country). Ethnicity did have a significant influence. African American families were more likely to be found in the average or above average clusters than in the below average cluster, compared to other ethnicities, χ² (2, n = 278) = 12.7, p < .01. Family beliefs are discussed as a possible explanation of this result.

Implications: The above average/average/below average continuum of family resilience protective factors supports theoretical concepts. Results also suggest a method for quantifying family resilience concepts. The unexpected absence of associations between protective factors, risks, and most outside supports, and the apparent resilience of African American families, despite lower SES and higher neighborhood problems and life events, suggest re-evaluating family resilience concepts or improving construct measurement. Measuring family resilience is important for developing and evaluating family-focused prevention programs.