Abstract
As many as 90% of persons with dementing illness demonstrate problem behaviors that range from repetitive verbalizations, agitation, and wandering to verbal and physical aggression toward self and others.Reliable and accurate measurement of these behaviors is crucial for tracking illness progression; for monitoring the effects of pharmacologic and behavioral interventions; and for continued investigation into the correlates of caregiver stress, burden, and coping. However, there is no single, universally accepted measure or methodology for operationalizing problem behaviors, and variations in definition and measurement across studies complicate drawing meaningful conclusions about these behaviors. This article is an overview of five factors that have complicated accurate and dependable measurement of problem behaviors in dementia: the shifting domain of problem behaviors, slippage across research constructs, unexplored rater bias, scoring bias, and the absence of benchmarking studies. A methodological agenda is discussed for future investigations in this rapidly growing area of gerontological research.