Energy, fatigue, or both? A bifactor modeling approach to the conceptualization and measurement of vitality
Vitality is an important domain reflecting both the physical and emotional components of health-related quality of life. Because of its complexity, it has been defined and measured both broadly and narrowly. We explored the dimensionality of a very comprehensive item bank hypothesized to measure vitality and its related concepts.
Secondary analyses were conducted using the responses of 1,343 adults representative of the US general population to Internet-based surveys including 42 items compiled from multiple scales (e.g., SF-36 Vitality, PROMIS-Fatigue), covering a broad range of vitality-related content areas (energy, fatigue, and their interference with physical, mental, social activities, and quality of life). Exploratory and confirmatory factor models were evaluated independently using split-half samples. Bifactor model was used to assess the essential unidimensionality of the items, in comparison with traditional unidimensional, multidimensional, and hierarchical models. Method effects of a common scale or phrase were modeled via correlating errors.
The exploratory factor analysis identified one dominant factor. The confirmatory factor analysis identified a best-fitting (CFI = 0.964, RMSEA = 0.084) bifactor model with one general (vitality) and two group (energy and fatigue) factors, explaining 69, 3, and 4 % of total variance. Correlating errors accounting for the method effects were important in identifying the substantive dimensionality of the items.
The bifactor model proved to be useful for evaluating the dimensionality of a complex construct. Results supported conceptualizing and measuring vitality as a unidimensional energy-fatigue construct. We encourage future studies comparing practical implications of measures based on the broader and narrower conceptualizations of vitality.