Abstract—In this paper, a Bayesian criterion-based method called the Lv measure, as well as its calibration distribution, is introduced and applied to model selection of structural equation models with ordered categorical data. A simulation study is presented to illustrate the satisfactory performance of the Lv measure in model selection. A quality of life data is analyzed for illustration.
Index Terms—ordered categorical data; Lv measure; Model selection; MCMC algorithm.
Yunxian Li is with the School of Finance, Yunnan University of Finance and Economics, Kunming, Yunnan, China. (phone: 86-15288336639;fax: 86-871-5113384; email: firstname.lastname@example.org).
Aijun Yang is with the School of Finance, Nanjing Audit University, Nanjing, Jiangsu, China. (e-mail: email@example.com).
Cite: Yunxian Li and Aijing Yang, "A Bayesian Criterion-based Statistic for Model Selection of Structural Equation Models with Ordered Categorical data," International Journal of Modeling and Optimization vol. 1, no. 2, pp. 151-157, 2011.