From a public health perspective, the measurement of cognitive function using representative population-based samples has gained increasing relevance in aging studies. However when doing so, several issues have to be considered. The purpose of this presentation is to provide a summary of these issues and illustrate the points with empirical results using longitudinal population-based surveys. The following aspects will be discussed: Valid definitions and methodological approaches need to be applied within a conceptual model that guides the analysis so that reliable and valid measures are used and population patterns can be identified. In addition, the use of direct and proxy assessments of cognitive function is common, and appropriate treatment of the two assessments is needed to establish population-based categories of cognitive functioning. Another common issue is the establishment of criteria for both cognitive domains and global cognitive assessments in order to establish categories of function such as normal functioning, cognitive impairment, and dementia. Deciding how to measure cognitive decline across longitudinal waves of data is another methodological challenge, in particular when assessment batteries change across survey waves.