Gravar-mail: Why generalized structured component analysis is not universally preferable to structural equation modeling