STRUCTURAL EQUATION MODELING - REPLY

In our previous article (B. S. Randhawa, J. E. Beamer, & 1. Lundberg, 1993), we concluded that (a) the endogenous latent variable (self-effi cacy) mediated the exogenous latent variable (attitude) and the endoge nous latent variable (mathematics achievement) and that (b) self-effic acy and attitude were more strongly related to achievement for boys th an for girls. H. W. Marsh et al. (1994) claimed that we inappropriatel y applied the structure equation modeling and consequently our conclus ions are unsubstantiated. We defend our analysis and conclusions. We s how that a philosophically different perspective in their reanalysis o f the covariance matrices and the fact that their input data were deri ved from our reported truncated summary results may have been the basi s for the differences in their results and conclusions.