Which time series test should a researcher chose to best describe the interactions among a set of time series variables? Aimed at providing social scientists with practical guidelines for identifying the appropriate multivariate time series model to use, this book explores the nature and application of these increasingly complex tests. Other topics it covers are joint stationarity, testing for cointegration, testing for Granger causality, and testing for model order, and forecast accuracy. Related models explained include transfer function, vector autoregression, error correction models, and others. Readers with a working knowledge of time series regression will find this helpful book accessible.

Taking a sequential approach to time-series model building, this book explores how to test for stationarity, normality, independence, linearity, model order, and properties of the residual process. The authors clearly define each testing procedure and offer examples to illustrate each concept. The authors also provide advice on how to perform the tests using different software packages. "This provides a nice roadmap for those doing time series analysis, and the authors should be applauded for this... Their approach is straightforward and logical and I believe will be useful many practicing statisticians." --Technometrics