Closed runnytone closed 6 years ago
It shouldn't be difficult to add. For now, nodeofactor("vertex.names")
and nodeifactor("vertex.names")
should produce the same results.
Thanks for your help. I find that nodeocovar and nodeicovar also work.
nodeicovar
and nodeocovar
are completely different in what they do, however: they are one-parameter statistics that are intended to represent positive correlation among values of edges incident on the same node.
More precisely, they also try to account for degree heterogeneity, but marginally rather than conditionally.
Fixed in statnet/ergm@7a67d2b16ae4e0e7410b2784ff52c83ab6bb575c.
Thanks for your contribution. Is there sender and receiver effect for valued ergm?