This chapter illustrates the systematic application of a multivariate perspective using SEM to explore a topic. In this presentation, the statistical details of the analyses will be ignored; these have been presented in earlier chapters, or can be found in the various publications referenced throughout. Here, the emphasis is on illustrating the broad enterprise of developing, evaluating, refining, and expanding multivariate models in order to understand system behavior and regulation. Throughout, the focus will be on the research enterprise rather than the analytical details. Thus, the philosophy and practice of SEM will be in the forefront, while the analysis of covariances, maximum likelihood, and mathematical details will be de-emphasized.

Background studies and findings

In 1992, Laura Gough and I conducted a study designed to examine the relationship between plant community biomass and species richness. This work was conducted in coastal marsh communities. The purpose of this study was to first characterize the relationship between biomass and richness. Then we planned to determine the role of competition in controlling the relationship. We expected that we would find a unimodal relationship between biomass and richness, primarily because of several key papers that had been published previously (Al-Mufti et al. 1977, Huston 1980, Wheeler and Giller 1982, Moore and Keddy 1989, Wisheu and Keddy 1989, Shipley et al. 1991, Wheeler and Shaw 1991). We also expected this relationship because there were several competing theories attempting to explain this phenomenon (Grime 1979, Huston 1979, Tilman 1982, Taylor et al. 1990, Keddy 1990).