Defining higher order dimensions of self-reported health: further evidence for a two-dimensional structure

We illustrate a method for understanding the extent to which a pool of fairly diverse health status questions can be examined for their dimensionality, by first fitting the data to a unidimensional measurement model, then examining item fit and residual factor analysis to determine the pattern of deviations from unidimensionality. We simultaneously administered five questionnaires to 1,714 people with cancer and HIV disease and combined all responses into one analysis. Item responses were cocalibrated onto one presumed underlying construct, often referred to as health-related quality of life (HRQOL). To identify significant, higher order substructures in the data, we then conducted a factor analysis of the data residuals, revealing two definable higher order dimensions: physical well-being and mental well-being. These results agree with other research and provide direction for further studies defining these dimensions within an item response theory (IRT) tradition.