Hi all, the function jtest allows to compare two non-nested models. The comparison is made such that the fitted values of one model are included in the regressor matrix of the other model. It then looks whether there is any predictive power of these fitted values. Unfortunately, the input has to come from a linear regression of class lm. I think that in the function jtest, both the regressor matrix of model 1 and the fitted values of model 2 will be estimated. Afterwards, it will perform the J test. However, I don't use lm and wanted therefore to ask if you know whether there is a way to manually define a regressor matrix (e.g. a matrix) and the fitted values (vector) as inputs? Best, Reiko -- View this message in context: http://r.789695.n4.nabble.com/Help-with-function-jtest-inputs-other-than-from-lm-tp4706738.html Sent from the R help mailing list archive at Nabble.com.