In recent years, epidemiologists have given increased attention to cognition, especially to the dementing illnesses that occur in old age. Central to this study is the measurement of change in cognition as opposed to cognition measured at a single point in time. This article addresses conceptual and methodological issues in the study of changes in cognitive function, including: 1) difficulties encountered with the use of single measurements of cognition and the importance of measuring changes in cognition; 2) sources of measurement variation and its potential effects; 3) the importance of careful modeling of age and education; 4) considerations in categorizing outcomes or combining the results of cognitive tests; and 5) the benefits of using multiple-outcome statistical models.