Design Considerations

The preceding chapters have largely been concerned with methods for analyzing and interpreting panel data that have already been collected. Here we turn our attention to considerations that arise in the data collection phase of research. We will discuss such questions as: Should one perform a panel study, or will some simpler research design suffice? How much time should separate the waves of a panel study? How can one minimize sample attrition, and adjust for attrition that has occurred? Only to a limited degree are these statistical questions. To a great extent they concern pragmatic questions, such as how to locate the subjects of a panel study in follow-ups. Our emphasis will be on strategies researchers have developed for dealing with such questions in their own work.