 |
 |
Contributions: Abstract
This workshop approaches SEM using the Bentler-Weeks model (a general and
simpler alternative to LISREL) and its implementation in EQS 6.1. Topics
include the practical (modeling with Diagrammer or text files) and theoretical
(use of recently published and in press statistics). Topics will be selected
from the following, depending on participant interest and background:
Typical models. Latent variables. Factor sc ores. Identification, estimation, testing, interpretation. Model modification: z, LM, an d Wald tests. Fit indices and power. Mean structure models. Growth curve models, and their automatic setup. Best practices with nonnormal data: Satorra-Bentler tests and Yuan-Ben tler residual-based statistics. Missing data estimators, tests, and SE's for nor mal (Jamshidian-Bentler) and nonnormal (Yuan-Bentler) data. Kim-Bentler tests of homogeneity of means and covariance matrices. Correlation structure methods for norm al and nonnormal data, and categorical variables. Simulation, bootstrap and jacknife methods. Case-weighting for complex samples and robust statistics (Yuan-Bentler) for outliers and influential points. Multi-sample/multi-group models. Multilevel models with ML (Ben tler-Liang) and robust tests and SEs (YuanBentler); HLM approaches. Other possible topics include internal consistency reliabilities: alpha, 1-factor, x-factor, arbitrary model, maximal, dimension-free, greatest lower bound.
Instructor
Peter M. Bentler is Professor of Psychology and Statistics at UCLA, where he works on the psychometric and statistical theory in SEM as well as on the etiology and consequences of drug abuse. He has been an elected president of the Society of Multivariate Experimental Psychology, the Psychometric Society, and the Division of Evaluation, Measurement, and Statistics of the American Psychological Association. In 1996, he received the Distinguished Scientific Contributions Award from the latter society.
Course Material
Participants will receive a CD that contains a time-limited version of EQS 6.1, along with pdf copies of articles from the recent literature as well as from Bentler, P. M., & Wu, E. J. C. (2002). EQS 6 for Windows User's Guide. Encino, CA: Multivariate Software, and Bentler, P. M. (2004). EQS 6 Structural Equations Program Manual. Encino, CA: Multivariate Software. More information on EQS 6.1 can be obtained at www.mvsoft.com.
Typical models. Latent variables. Factor sc ores. Identification, estimation, testing, interpretation. Model modification: z, LM, an d Wald tests. Fit indices and power. Mean structure models. Growth curve models, and their automatic setup. Best practices with nonnormal data: Satorra-Bentler tests and Yuan-Ben tler residual-based statistics. Missing data estimators, tests, and SE's for nor mal (Jamshidian-Bentler) and nonnormal (Yuan-Bentler) data. Kim-Bentler tests of homogeneity of means and covariance matrices. Correlation structure methods for norm al and nonnormal data, and categorical variables. Simulation, bootstrap and jacknife methods. Case-weighting for complex samples and robust statistics (Yuan-Bentler) for outliers and influential points. Multi-sample/multi-group models. Multilevel models with ML (Ben tler-Liang) and robust tests and SEs (YuanBentler); HLM approaches. Other possible topics include internal consistency reliabilities: alpha, 1-factor, x-factor, arbitrary model, maximal, dimension-free, greatest lower bound.
Instructor
Peter M. Bentler is Professor of Psychology and Statistics at UCLA, where he works on the psychometric and statistical theory in SEM as well as on the etiology and consequences of drug abuse. He has been an elected president of the Society of Multivariate Experimental Psychology, the Psychometric Society, and the Division of Evaluation, Measurement, and Statistics of the American Psychological Association. In 1996, he received the Distinguished Scientific Contributions Award from the latter society.
Course Material
Participants will receive a CD that contains a time-limited version of EQS 6.1, along with pdf copies of articles from the recent literature as well as from Bentler, P. M., & Wu, E. J. C. (2002). EQS 6 for Windows User's Guide. Encino, CA: Multivariate Software, and Bentler, P. M. (2004). EQS 6 Structural Equations Program Manual. Encino, CA: Multivariate Software. More information on EQS 6.1 can be obtained at www.mvsoft.com.