A Note on Heywood Cases in Linear Structural Models

This Paper Shows That Rindskopf's Parameterization Technique Is Not an Acceptable Solution to the Heywood Problem, Especially, If One Is Interested in Standard Errors and T-Statistics for the Parameter Estimates of a Linear Structural Model. Simple Alternative Solutions Are Proposed and Discussed. Finally, This Note Raises the Question of the Existence of a Fundamental Incompatibility Between a Given Data Set and a Lisrel Modelling of a Problem. Keywords: Heywood Cases, Lisrel.