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

What I often do with ego-network data is to input it as it were a single
giant network (that is highly disconnected). For example, if you have
sampled 100 respondents (egos) and each person mentions 9 people, then you
have a network of 1000 people. Then you run ego net measures such as
structural holes measures on that one network. Seconds later, you have the
results for all nodes in the dataset, including your egos (and only the ones
for the egos are valid). 

A few notes: It is helpful to maintain a node attribute variable that is 1
if a node is an ego and 0 if not. This makes it easy to extract the right
numbers from the results. Also, you must assign unique id codes to all
nodes, including both egos and alters. For example, you might use an id
system in which egos are prefixed by "E" and alters' ids contain the id of
the ego they belong to. Here is a sample node list type data entry showing
the id codes:

E001 A001-1 A001-2 A001-3
A001-1 A001-3
A001-2 A001-1 A001-3 
E002 A002-1 A002-2 A002-3 A002-4

The first line shows ego 001's three alters. The 2nd and 3rd lines show ties
among the alters of ego 001. Note that it could be that alter A002-3 ("bill
smith") happens to be the same person as ego E001, but since this is ego net
data, you probably don't know it and definitely don't care.


Steve Borgatti
Professor & Chair, Organization Studies Dept.
Boston College
Tel: +1 978 456 7356; Fax: +1 978 456 7373
E-mail: [log in to unmask]
>-----Original Message-----
>From: Social Networks Discussion Forum [mailto:[log in to unmask]] On
>Behalf Of Daniele Mascia
>Sent: Monday, December 19, 2005 7:43 AM
>To: [log in to unmask]
>Subject: [SOCNET] how to treat ego-network data?
>*****  To join INSNA, visit  *****
>Dear Socnetters,
>we are conducing an analysis centered on the ego-networks of scientists
>working in several international scientific institutions. Our idea is that
>of investigating structural characteristics of the one-step neighborhood of
>the surveyed scientists.
>To date, we have found a few articles explaining concepts and
>charateristics related to the ego-network, but we are still in search of
>materials about the methodology to be used to deal with ego-network data.
>In particular, given a set of ego-networks collected in spreadshit editors,
>we would like to know which format is reccomended for importing such data
>in one of the available software for network analysis (e.g. UCINET,
>Pajek...), and how to conduce relative analyses.
>Thanks in advance for any comments.
>Best regards,
>Daniele Mascia
>Catholic University - Rome
>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.

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.