Back to Search Start Over

WHAT TETRAD CAN DO

Authors :
Richard Scheines
Clark Glymour
Peter Spirtes
Kevin T. Kelly
Publication Year :
1987
Publisher :
Elsevier, 1987.

Abstract

This chapter discusses the strengths and limitations of the TETRAD program. It presents a problem using simulated data of TETRAD model. On real data sets, TETRAD sometimes yields results that suggest that no linear model will adequately explain the data. In other cases, the program gives results that suggest that too many linear models will explain the data. In a study explained in the chapter data was generated using SYSTAT BASIC. The simulated data was produced for the exogenous variables with a pseudo-random-number generator distributed normally with mean 0 and variance 1. All other variables were linear combinations of their respective immediate ancestors in the directed graph of the model. The linear coefficients were also chosen at random, although their values for each particular model are non-stochastic constants. Kohn's Class and Conformity describes several large survey studies that investigate the relationships between social class, attitudes, and personality structure.

Details

Database :
OpenAIRE
Accession number :
edsair.doi...........c701a098faf59e1d490dacf032dc5f38
Full Text :
https://doi.org/10.1016/b978-0-12-286961-7.50012-3