Dear all,

  I realized that WLS estimation results in gretl are different from those in Stata 

  in the case when observations with weight=0 do exist.

  Now, in gretl, the mechanics of the algorithm drops these observations 

  from the estimation process. And indeed, both Stata and gretl report the same 

  number of observations (ie. observations actually involved in computing) with gretl 

  giving addtional info on dropped observations.

  Consequently, gretl doesn't provide postestimation values on these dropped 

  observations (uhat, yhat)...but Stata does!!! And I struggle to understand the logic

  behind Stata's proceeding and how it affects the estimates...

I don't know what Stata does there (although I suspect it is documented), but in principle it seems easy enough to compute a fitted value yhat even for any obs (y-X-constellation) which wasn't part of the estimation sample. I mean that's what we do for prediction, no? Does a (static) "fcast" work there in gretl?

thanks

sven