Abstract
Data which are measured in several characteristics and at several different points of time are sometimes available for a population. Transition matrices or mobility tables are often constructed from such ‘panel data’ and, as in the case of the Census and migration analysis, become the basis for a wide range of research activity. Problems arise when gaps appear in the series or when change needs to be tabulated over a specific interval. Several methods for estimating subinterval transition matrices are presented and evaluated, and one method is examined in an empirical situation. This is seen as a step towards full control and understanding of spatial and temporal-aggregation problems in data analysis.
Get full access to this article
View all access options for this article.