*****  To join INSNA, visit  *****


> I have a similar problem. In my networks, by design each person has a
> fixed (or very close) number of connections. That means no main
> effects of any covariate will be significant. So our interested is in
> segregation in friendship networks. Indeed, we find almost no
> covariates has significant predictive effect, 

If the degrees are fixed or otherwise constrained by the sampling
design, it's probably better to fit a model with the sample space
constrained to the networks that could have been observed under the
design. (In ergmese, that's the constraints=~bd(), with appropriate
arguments to bd().)

Modelling group heterogeneity using ERGM terms on top of that may or may
not be necessary, depending on how strict a constraint the design


Dr. Pavel N. Krivitsky
Lecturer in Statistics
Building 39C, Room 185
School of Mathematics and Applied Statistics
University of Wollongong, NSW 2522
Phone: +61 (2) 4221 3713

SOCNET is a service of INSNA, the professional association for social
network researchers ( To unsubscribe, send
an email message to [log in to unmask] containing the line
UNSUBSCRIBE SOCNET in the body of the message.