On Fraas and Newman's Goodness-of-Fit Test for Structural Equation Models

Fraas and Newman's (1994) binomial test of goodness of fit of structural equation models is based on the assumption that counts of path estimates consistent with a researcher's expectations can be viewed as, binomial random variables. In this article I argue that these counts do not have binomial distributions because of the lack of independence among estimates of path coefficients of structural equation models.