초록
영어
Structural Equation Modeling(SEM) has well been adopted and ever expanded in SLA to address increasing complexity of research questions. While the conventional covariance-based SEM is still popular and firmly established as a family of statistical techniques, it suffers from its inherent problems such as nonnormal data handling and unrealistic assumptions about the relationship of variables. Recently a few alternatives to the conventional SEM have emerged in the SEM literature. This study applied Bayesian SEM to modeling the L2 motivational self system and showed its inherent advantages over the conventional covariance-based SEM in comparison. It also discussed some methodological issues in applying the Bayesian approach to research in SLA.
목차
1. Introduction
2. Background theories
2.1. L2 motivational self system
2.2. Brief overview of Bayesian Inference
3. Methods
3.1. Instruments and participants
3.2. Modeling procedures
4. Results
5. Discussion
References
Appendix