Analyzing time budgets with Markov chains

Time budgets are generally estimated with data obtained by sampling fragments of animal activity. Theseestimates are often wrong due to nonrandom missing data, improper weighting of the observations, and bias from nonrandom sampling of initial behavior in an observation bout (pickup bias). Time budIget studies also suffer from inadequate statistical analysis in comparisons of groups. Markov chain models are discussed as a solution to these sampling and analysis issues. A description of the methods used to construct and analyze these models is provided, highlighted by a simplified beaver time budget example.