Dear StataListers, what should I do if I get a negative error variance in the stimation of structural equation model or in a confirmatory factor analysis? Which are the causes of this? Can you recomended some paper which talk about this? Thank you