Abstract:

Antedependence (AD) models -- also known as transition or Markov models -- are a useful, though not widely known, class of models for the covariance structure of longitudinal data.  Like stationary autoregressive models, AD models allow for serial correlation within subjects, but they are more general in the sense that they do not stipulate that the variance is constant over time nor that correlations between measurements equidistant in time are equal.  Thus, AD models provide a more parsimonious approach to the analysis of nonstationary data than the completely unstructured classical multivariate approach.  They also can provide important insights into the nature of the dependence structure of the data, which are not provided by a marginal or random-effects-based analysis. In this talk I will describe these models and methods for various estimation and testing problems associated with them.

The methods will be illustrated using data on cattle growth and long-distance running.