Recent advances in psychometric methods for categorical and other non- normal data, such as factor analysis, structural equation modeling, and longitudinal modeling. have considerable promise for applications In the fields of epidemiological and clinical research where they are not currently in common use. The major objective of this proposal Is to determine how best to tailor these methods for epidemiological and clinical applications. Statistical techniques will be developed and their sampling behavior investigated by simulated data. Real data will be used to try out the methodology in practice. The results of this research would be the basis for the development of computer software and/or pedagogical writings in Phase Il. Because epidemiological and clinical research is carried out not only by government agencies but also by universities and pharmaceutical companies around the world, there Is always a demand for new ways of better utilizing data in these research settings. Even the lack of software for and pedagogical writings about psychometrIc methods applied to epidemiology, it would appear there would be a demand for Phase II products.
