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

Hi all,

I have a dataset of inter-organizational collaborations.  However, because
I was unable to collect data for the entire organizational sample what I
essentially have is a 2-mode network of 22 designator organizations (i.e.,
organizations who took the survey) and 107 potential collaborator
organizations from which they could choose.  My primary research question
is basically, "What are the generative mechanisms, both endogenous and
exogenous, that predict being designated as a collaborator?"  My intent is
to use BPNet for the analysis, but I'm open to other options should
responses to my question below warrant it.

*My question to the group is*: Is there a way to include relational
attributes between actors in Group A (those who designated collaborators)
and organizations in Group  B (those named as collaborators)?

Obviously in 1-mode p*/ERGM analysis you can include dyadic relational
atributes. But as fas as I can tell, the same option doesn't seem to exist
for 2-mode analysis, at least not in BPNet (unless I'm missing something
obvious).  Examples of some of the relational variables I would like to
include in the model are: (1) communication frequency between both parties;
(2) a measure of patron similarity (i.e., the designator org. serves
similar patrons as the collaborator it named); and (3) designators'
perceptions of a specific collaborator's expertise.

It seems like the combined attribute parameter option in BPNet might allow
me to model effects of things like issue homophily or heterophily as a
predictor in the model (e.g., when a designator and a collaborator focus on
the same issue).  But I'm not sure whether or how I can include other types
of edge attributes, like frequency of interaction.

I realize that part of the problem may simply be my less than ideal dataset
and the slightly unconventional way of reconfiguring these relationships as
"affiliations".   However, I am eager to hear of suggestions or work-around
options so that I can analyze what I do have in a probabilistic manner.

Thanks in advance,

Lindsay Young
Doctoral Student
Northwestern University
Communication Studies
Media, Technology & Society Program

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.