ABSTRACT

The justification for using linear structural equation models with unmeasured variables, or LISREL (Linear Structural Relationships), can be found in the sociological literature addressing measurement error in causal models (Blalock [6,9]; Costner [16]; Siegal and Hodge [42]). Such techniques permit the construction of causal models that incorporate and correct for the effects of measurement error and are, therefore, fundamental in adequately estimating the parameters of causal models and assessing the goodness of fit of the model to the data.