Nonlinear Generalized Structured Component Analysis

Generalized Structured Component Analysis (GSCA) represents component-based structural equation modeling. Currently, GSCA is geared only for the analysis of quantitative data. In this paper, GSCA is extended to deal with qualitative data through data transformation. In particular, the optimal scaling approach is adopted for data transformation as it can be readily coupled with the GSCA estimation procedure. An alternating least-squares algorithm is developed that involves two phases for estimation of model and data parameters. Two empirical applications are presented to demonstrate the usefulness of the proposed method.