Dementia is defined as global cognitive impairment that interferes with func-tion; however, function has been less well measured than cognition in anti-dementia drug trials. In the modern era of anti-dementia clinical trials, measurement of function has improved by differentiating between the aspects of function that have been impaired – for example, impaired initiative versus ineffective performance, as is evaluated by the Disability Assessment for Dementia. Obstacles remain, including how best to distinguish the concepts of functional impairment and disability, how broad to make the range of functional impairment (e.g. whether it should include impaired performance of hobbies, or withdrawal from leisure activities) to individualize assessment, and to distinguish cognitive from non-cognitive causes of impaired function. Even though it appears that improved function is commonly related to improved executive function, and that the latter, more than other aspects of improved cognition, is the effect most prized by patients, families and physicians, functional assessment measures are often less sensitive to change than other outcome measures. How best to improve sensitivity to change of functional measurements is controversial, but it is necessary to do so in order to evaluate the full effects of treatment.