Consolidated measures of activity among older adults: results of a three data set comparison.

This study explores the potential to consolidate a broad range of activity items to create more manageable measures that could be used in statistical modeling of multi-activity engagement. We utilized three datasets in the United States: Panel Study of Income Dynamics, Health and Retirement Study, and Midlife in the United States. After identifying activity items, exploratory and confirmatory factor analysis were used to empirically explore composite activity measures. Findings suggest that discrete activity items can be consolidated into activity domains; however, activity domains differ across datasets depending on availability of activity items. Implications for research and practice are further discussed.