Dear Statalist members,

I have an identical problem as the one previously published in the forum: https://www.stata.com/statalist/arch.../msg01272.html

Using the command spatgsa to obtain Moran's I, the output had no results. One argument is that I have 5 counties without observations.

But, if I eliminate those ones, then I get the following error message:

Matrix NW2 is 28x28, the dataset in use has 23 obs.
To run -spatgsa- weights matrix dimension must equal N. of obs
What should you recommend as possible approach?

Thank you in advance.