Using TETRAD II as an Automated Exploratory Tool

Abstract

An automated process for forming measurement models and structural equation models was developed using an experimental version of TETRAD II. An evaluation function was developed to select the measurement models and the structural models for further consideration. Comparisons were made between developing the models in this way and using factor analysis to develop measurement models. We found TETRAD II useful in building sets of measurement models and in suggesting sets of possible structural models. A previously unstudied data set with a sample size of 7,625 was used for testing the process. We developed measurement models and structural equation models that satisfied our intuitive understanding and had a chi-square goodness-of-fit p value of .77. Keywords: EQS, structural models, measurement models, model specification, TETRAD.

Get full access to this article

View all access options for this article.

References

Bollen, K. 1989. Structural equations with latent variables. New York: Wiley.

Chambers, J. , Cleveland, W. , Klemer, B. , & Tukey, P. 1983. Graphical methods for data analysis. Belmont, CA: Wadsworth.

Cook, R. , & Weisberg, S. 1990. Added variable plots m linear regression. In Proceedings of IMA Conference on Robustness and Diagnostics. New York: Springer-Verlag.

Draper, N. , & Smith, H. 1981. Applied regression analysis. 2d ed. New York: Wiley.

Edwards, D. , & Havránek, T. 1985. A fast procedure for model search in multidimensional contingency tables. Biometrika 72:339-51.

—. 1987. A fast model selection procedure for large families of models. Journal of the American Statistical Association 82:205-13.

Efron, B. , & Tibshirani, R. 1991. Statistical data analysis m the computer age. Science 253:390-95.

Glymour, C. , Schemes, R. , Spirtes, P. , & Kelly, K. 1987. Discovering causal structure. San Diego, CA: Academic Press.

Glymour, C. , & Spirtes, P. 1988. Latent variables, causal models, and overidentifying constraints. Journal of Econometrics 39:175-98.

10.

Lehmann, E. 1990. Model specification: The views of Fisher and Neyman, and later developments. Statistical Science 5:160-68.

11.

Schemes, R. , Spirtes, P. , & Glymour, C. 1991. Building latent variable models. Tech. rep. LCL-91-2, Carnegie Mellon University.

12.

Spirtes, P. 1989. A necessary and sufficient condition for conditional independencies to imply a vanishing tetrad difference. Tech. Rep. CMU-LCL-89-3, Laboratory for Computational Linguistics, Carnegie Mellon University.

13.

Spirtes, P. , & Glymour, C. 1991. An algorithm for fast recovery of sparse causal graphs . Social Science Computer Review 9:62-72.

14.

Spirtes, P. , Schemes, R. , & Glymour, C. 1990 Simulation studies of the reliability of computer-aided model specification using the TETRAD II, EQS, and LISREL programs. Sociological Methods and Research 19:3-67.

15.

—. In press. Causality, prediction, and search. Lecture Notes in Statistics. New York: Springer-Verlag.

16.

Weihs, C. , & Schmidli, H. 1990. OMEGA (Online multivariate exploratory graphical analysis): Routine searching for structure. Statistical Science 5:175-226.

17.

CART. California Statistical Software, Inc., 961 Yorkshire Ct., Lafayette, CA 94549; 415-283-3392.

18.

EQS. Available from BMDP Statistical Software, Inc., 1440 Sepulveda Blvd., #316, Los Angeles, CA 90025; 213-479-7799.

19.

SAS. SAS Institute, POB 8000, Cary, NC 27512; 919-467-8000.

20.

SPSS/PC + CHAID. SPSS, Inc., 444 N. Michigan Avenue, Chicago, IL 60611; 312-329-3300.