Structural Equation Modeling and Natural SystemsThis book, first published in 2006, 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. |
Contents
| 326 | |
Section 2 | 332 |
Section 3 | 333 |
Section 4 | 337 |
Section 5 | 339 |
Section 6 | 340 |
Section 7 | 345 |
Section 8 | 346 |
Section 9 | 347 |
Section 10 | 348 |
Other editions - View all
Common terms and phrases
bird’s cluster variable composite model composite variable correlations COVARIANCE MATRIX data in Table degrees of freedom differ among groups Discussion of Illustration effects of x1–x equal across groups equal x2 ESTIMATION TERMINATED NORMALLY ETA2 example Figure A.1 FIT Chi-Square Test Fit Value Degrees Freedom P-Value habitat Illustration 1.3 interest interpret latent variable models LISREL Mean Square Error mediated by y1 model chi-square MODEL FIT Chi-Square Model Fit Value model in Figure Mplus multigroup analysis omits the paths output parameter estimates path from x2 paths with question population densities possible applications presented problem recode represent reproductive success response variable Results for Illustration Root Mean Square sample situation stage of analysis standard deviations standardized coefficients standardized path coefficients Table A.1 TESTS OF MODEL unmeasured unstandardized parameters variance of y2 variance–covariance matrix x1 and x2 x2 to y2 Y1 and Y2 y2 in group Y3 and Y4 yy22 yy33 Figure