Refining quality of life: validating a multidimensional factor measure in the severe mentally ill.
Quality of life measurement has historically been characterized by a focus on physical functionality, great variability in definitions and insufficient attention to psychometric properties of measures. The present study examined four core subscales of the Quality of Life Enjoyment Scale (Q-LES-Q) designed to assess subjective quality of life (i.e. physical health, subjective feelings, leisure activities and social relationships) administered to 151 male and female subjects with severe mental illness admitted to a residential community treatment center or a university psychiatric hospital. The use of factor analysis is a common approach to examining construct validity of instruments through the examination of correlated clusters of item responses. Those sets of highly correlated item responses should identify a construct or dimension of related items (i.e. a factor). Two factor approaches, exploratory (i.e. the maximum number of possible factor is unspecified) and restricted (i.e. the maximum number of factors allowed is specified), were employed to examine construct validity of the four Q-LES-Q subscales. In addition, both orthogonal (i.e. independence between factors is maximized) and oblique (i.e. correlated factors are allowed) rotations (arrangements) of factor structure were also investigated to define subscale validity further. Results supported good construct validity for each subscale with either factor approach, i.e. the four proposed subscales were clearly identifiable in the factors (groupings) of correlated item responses from the sample. In both cases, the orthogonal (independent) rotation produced the simplest structure, i.e. the clearest groupings of items. These results indicate that the Q-LES-Q does appear to measure valid dimensions matching those proposed by the subscales and, thus, may be a useful and reliable tool for clinical applications.