Hi all,

I've made a matrix based on the similarity of organizations in forums (Euclidean distance): fairly dense, little zeroes. When I spmatrix with spectral normalization, Stata reads the matrix as having ALL islands. This is not true. When I row or min normalize, I do not have an error. I've used spectral normalization on numerous other matrices calculated the same way and have not had an error. I'm at a loss. Any ideas would be greatly appreciated.

Thanks!