
Full text loading...
Common features of longitudinal surveys are complex sampling designs, which must be maintained and extended over time; measurement errors, including memory errors; panel conditioning or time-in-sample effects; and dropout or attrition. In the analysis of longitudinal survey data, both the theory of complex samples and the theory of longitudinal data analysis must be combined. This article reviews the purposes of longitudinal surveys and the kinds of analyses that are commonly used to address the questions these surveys are designed to answer. In it, I discuss approaches to incorporating the complex designs in inference, as well as the complications introduced by time-in-sample effects and by nonignorable attrition. I also outline the use and limitations of longitudinal survey data in supporting causal inference and conclude with some summary remarks.
Article metrics loading...
Full text loading...
Data & Media loading...