The book "Latent Variable Path Modeling with Partial Least Squares" by Jan-Bernd Lohmöller is a comprehensive treatise on the application and statistical foundation of Partial Least Squares (PLS) in the context of latent variable path modeling. The author,深受多位教授的指导和启发，自1978年首次接触Herman Wold的PLS方法以来，对这一方法产生了浓厚的兴趣。本书的核心内容分为几个部分： 1. **Basic Principles of Model Building**: This section introduces the fundamental concepts of model building, distinguishing between empirical and theoretical concepts, causation and prediction, and the differences between data modeling and covariance modeling. 2. **The Basic and the Extended PLS Method**: Here, the author describes the basic and extended PLS methods, including model specification, the basic PLS algorithm, and extensions such as the Split Principal Component theorem. The section also covers estimation methods and the assessment of results. 3. **Foundations of Partial Least Squares**: This part delves into the statistical foundations of PLS, including conditional expectation, predictor specification, principal components, factor score estimation, and predictive two-block models. It also introduces the Split Principal Components and Split Multiple Regression concepts. 4. **Mixed Measurement Level Multivariate Data**: This section addresses the handling of mixed measurement levels in multivariate data, including categorical variables and contingency tables. It provides methods for analyzing such data using PLS. 5. **Predictive vs. Structural Modeling: PLS vs. ML**: This chapter compares PLS with the Maximum Likelihood (ML) estimation method, discussing the differences between scored and unscored latent variables, consistency, and bias in two-block models. 6. **Latent Variables Three-Mode Path (LVP3) Analysis**: This part introduces three-mode data models and the Kronecker Principal Component (KPC) model, along with the three-mode LVP (LVP3) model. It also covers the PLS estimation of LVP3 models and applications in longitudinal data. 7. **PLS Programs and Applications**: This section provides an overview of PLS programs and their applications in various fields, including psychological and educational research. The book is structured to be both mathematically rigorous and applied, making it a valuable resource for researchers and practitioners in the field of latent variable modeling.The book "Latent Variable Path Modeling with Partial Least Squares" by Jan-Bernd Lohmöller is a comprehensive treatise on the application and statistical foundation of Partial Least Squares (PLS) in the context of latent variable path modeling. The author,深受多位教授的指导和启发，自1978年首次接触Herman Wold的PLS方法以来，对这一方法产生了浓厚的兴趣。本书的核心内容分为几个部分： 1. **Basic Principles of Model Building**: This section introduces the fundamental concepts of model building, distinguishing between empirical and theoretical concepts, causation and prediction, and the differences between data modeling and covariance modeling. 2. **The Basic and the Extended PLS Method**: Here, the author describes the basic and extended PLS methods, including model specification, the basic PLS algorithm, and extensions such as the Split Principal Component theorem. The section also covers estimation methods and the assessment of results. 3. **Foundations of Partial Least Squares**: This part delves into the statistical foundations of PLS, including conditional expectation, predictor specification, principal components, factor score estimation, and predictive two-block models. It also introduces the Split Principal Components and Split Multiple Regression concepts. 4. **Mixed Measurement Level Multivariate Data**: This section addresses the handling of mixed measurement levels in multivariate data, including categorical variables and contingency tables. It provides methods for analyzing such data using PLS. 5. **Predictive vs. Structural Modeling: PLS vs. ML**: This chapter compares PLS with the Maximum Likelihood (ML) estimation method, discussing the differences between scored and unscored latent variables, consistency, and bias in two-block models. 6. **Latent Variables Three-Mode Path (LVP3) Analysis**: This part introduces three-mode data models and the Kronecker Principal Component (KPC) model, along with the three-mode LVP (LVP3) model. It also covers the PLS estimation of LVP3 models and applications in longitudinal data. 7. **PLS Programs and Applications**: This section provides an overview of PLS programs and their applications in various fields, including psychological and educational research. The book is structured to be both mathematically rigorous and applied, making it a valuable resource for researchers and practitioners in the field of latent variable modeling.