Alzheimer’s disease (AD) is a chronic disease of moderate to long duration. Numerous studies have used variations in the rate of progression of AD to compare therapies and the effects of various bio-medical markers and for clinical use. Studying progression is more difficult if the measure of severity does not change linearly, since the random effects models are difficult to interpret if there are nonlinear terms. Preliminary examinations suggest that the Dementia Severity Rating Scale (DSRS) has great potential in this regard. The DSRS is a measure of behavioral impairment designed for measuring severity in AD. It appears that the DSRS increases linearly with duration of disease until the most severe stages. In a small proportion of patients, the DSRS remains constant at a low value for a short period of time before beginning a linear trend. The aims of the proposed research are to demonstrate the pattern of change in the DSRS and the variation in progression rates among cases and to calibrate the DSRS by examining correlates of progression rates and average DSRS scores at the time of major events such as placement in a care facility or death.