A multilevel analysis of whole family functioning using the McMaster Family Assessment Device.

This study demonstrates the use of multilevel modeling to examine influences on ratings of whole family functioning collected from multiple family members (N=26,614) living in 11,023 families with 1 or more dependent children aged 0 to 24 years. Results indicate that 45.7% of the variance in ratings of whole family functioning was shared among family members, whereas 54.3% was nonshared. Family-level characteristics, such as socioeconomic status (SES), family structure and composition, and family well-being, accounted for 30.6% of between-family variation (i.e., shared perceptions). Individual-level characteristics, such as sex, age, dependent child status, education, and well-being, accounted for 5.6% of within-family variance (i.e., unique perceptions). There was significant between-family variation in the relationship between dependent child status and ratings of family functioning, and increased rating discrepancies among members of the same family were linked with higher levels of family SES. The findings attest to the validity of measuring whole family functioning directly from self-report ratings provided from multiple family members. However, caution is warranted when assessments are available only from single respondents. (c) 2008 APA, all rights reserved