Factor analysis of nested data is a challenge for researchers when they need to accurately identify the most appropriate latent configuration of self-report instruments. The present study used a multilevel framework to evaluate the factor structure underlying the 12-item three-factor Family Relationships Index (FRI), while adapting this instrument to the Italian context. By way of separating the two sources of variance (within and between families), results from 231 family members nested in 77 family triads supported a three-factor model (i.e., family cohesion, communication, and conflict resolution). Multilevel confirmatory factor analysis (MCFA) corroborated this model at the family level in particular. A one-factor model was also tested but resulted in being less suitable at both the individual (within) and family (between) level of analysis. Finally, we consider challenges in using such statistical techniques, while taking into account the FRI’s briefness and easiness to complete and score it in a practice-oriented assessment perspective.
- family assessment
- multilevel confirmatory factor analysis