Page 1 of 1

partitioned dataset won't run EGLS

Posted: Wed Oct 30, 2019 11:26 am
by eljayski
Hello. I have a data set that runs EGLS just fine, (280 cross section units and mostly 8 years but unbalanced).

I've eliminated a large number of observations pursuant to a research objective and also removed observations representing only a single year.

Model no longer runs with reduced data and I get this message: "Positive or non-negative argument to function expected in
computation of group weight (variance)."

What might I be missing?

Many thanks, eljayski