Subject: Abstract:Techniques for analyzing vertebrate social structure (fwd)

Mike Williamson (
Sat, 13 Mar 1999 09:28:52 -0500 (EST)

---------- Forwarded message ----------
Date: Fri, 12 Mar 1999 09:16:36 -0800
From: MARMAM Editors <marmamed@UVic.CA>
Reply-To: Marine Mammals Research and Conservation Discussion
Subject: Abstract:Techniques for analyzing vertebrate social structure              (fwd)

The following paper may be of interest to those studying cetacean (and
other) social organizations.

Hal Whitehead (

Whitehead, H. and S. Dufault.  1999. Techniques for Analyzing Vertebrate
Social Structure Using Identified Individuals: Review and Recommendations.
Advances in the Study of Behavior 28: 33-74.


 The objective of this chapter has been to help ethologists choose
 appropriate methods for the analysis of vertebrate social structures.
  We examined 88 studies of non-human vertebrate social structure in
 which animals were individually identified and interactions or
 associations between pairs of animals were recorded.  Most studies
 were of 'fission-fusion' type societies and concerned mammals,
 especially ungulates, primates and cetaceans.  In these studies the
 fundamental data consisted of observations of behavioral
 `interactions', or `associations' defined by spatial or temporal
 proximity, and/or presence in the same group.  Data from different
 sampling periods were then combined into one or more measures of
 `relationship' for each pair of animals.  Most frequently, just one
 measure of relationship was calculated, an `association index'.
 There are difficulties in choosing a suitable association index.  To
 represent the social structure of the population, matrices of
 association indices, or other measures of relationship, were
 displayed using cluster analyses, sociograms, multidimensional
 scaling and other techniques.  The temporal patterning of pairwise
 interactions or associations, an important element of relationships,
 and thus of social structure, generally received only superficial
 treatment.  Among our recommendations for future studies are: records
 of interactions are preferable to those of associations; when used,
 associations should be defined on the basis of the likelihood of
 interaction; association can be determined by presence in the same
 group but groups should be defined so that most interactions take
 place within them; the temporal patterning of interactions or
 associations should be an important part of such analyses; a
 multivariate representation of several measures of relationship may
 be useful; and it is often instructive to compare the distribution of
 measures of relationship, or some other statistic, with that expected
 from 'null' models of social structure.

-To submit a message to MARMAM, send it to:
-Please include your name and e-mail address in the body of the
text of all submissions, and ensure your message has an appropriate
subject heading (ie., not "Message for MARMAM")
-To subscribe to MARMAM, send a message to:
saying: subscribe marmam Yourfirstname Yourlastname
-To contact the MARMAM editors, write to:
-MARMAM Editorial Policy & FAQ: