Structural Equation Modeling (SEM) is a technique that is used to estimate, analyze and test models that specify relationships among variables. This book explains the theory behind the statistical methodology, including chapters on conceptual issues, the implementation of an SEM study, and the history of the development of SEM. It provides examples of analyses on biological data including multi-group models, means models, p-technique and time-series. In addition, the book discusses computer applications and contrasts three popular SEM software packages. Data sets and programs in the book can be downloaded from http://nrmsc.usgs.gov/products/Pugesek_SEM.htm.

This book presents an introduction to the methodology of structural equation modeling, illustrates its use, and goes on to argue that it has revolutionary implications for the study of natural systems. A major theme of this book is that we have, up to this point, attempted to study systems primarily using methods (such as the univariate model) that were designed only for considering individual processes. Understanding systems requires the capacity to examine simultaneous influences and responses. Structural equation modeling (SEM) has such capabilities. It also possesses many other traits that add strength to its utility as a means of making scientific progress. In light of the capabilities of SEM, it can be argued that much of ecological theory is currently locked in an immature state that impairs its relevance. It is further argued that the principles of SEM are capable of leading to the development and evaluation of multivariate theories of the sort vitally needed for the conservation of natural systems. Supplementary information can be found at the authors website, http: //www.jamesbgrace.com/. Details why multivariate analyses should be used to study ecological systems Exposes unappreciated weakness in many current popular analyses Emphasises the future methodological developments needed to advance our understanding of ecological systems