***** To join INSNA, visit http://www.insna.org *****
Recently, a message was posted on SOCNET concerning "The Social
Organization of Conspiracy" (Baker and Faulkner 1993). Subsequently, the
sender of this post contacted me via email. I sent a response to him, but
since he raised concerns in this public forum, I thought it appropriate to
post my response. The relevant portions are reproduced below.
The confusion is caused by a typo in the article, and a little lack of
clarity on our part. Otherwise, our measures, analysis, findings, and
interpretation are correct. Specifically, here are the answers to the
issues you raised:
(1) Our definition of closeness (farness) point centrality, Eq. 3, p.
848. There is a typo. The superscript -1 is missing from the equation,
which, unfortunately, we didn't spot in proof reading. This is a measure
of "point decentrality" (Freeman 1979:225), or, as we put it, farness.
(2) Comparisons of point centrality across networks of different sizes. It
make sense to standardize measures of point centrality in most cases--but
not all (Freeman 1979). In our context, degree centrality has a very
intuitive meaning: the number of direct eyewitnesses of ego's
participation in a price-fixing conspiracy. In a legal setting, the number
of direct eyewitnesses (degree centrality) makes more sense than the number
of direct eyewitnesses relative to the total number of participants
(relative degree centrality). A direct eyewitness is a direct eyewitness,
whether the network has 21 participants (transformers), 24 (turbines), or
33 (switchgear). (Indeed, it would be hard to say how a grand jury would
interpret relative degree centrality.) By extension, the same logic
applies to farness and betweenness. Farness, for example, measures paths
(geodesics) of direct and indirect eyewitnesses. So, in our particular
context, it makes sense to *not* standardize point centrality measures.
(3) For the measure of graph centralization, we used the standardized
closeness index, according to Eq. 4, p. 850. We did not use farness. We
apparently caused confusion by referring to farness when discussing graph
centralization (e.g., footnote 8). We did so to remind the reader that our
measure of "closeness" point centrality was actually a measure of
distance. In this context, it seemed especially important to do so. In
retrospect, we could have been clearer about this in our article, making
sure there was no ambiguity about our measures of point centrality and
SOCNET is a service of INSNA, the professional association for social
network researchers (http://www.insna.org). To unsubscribe, send
an email message to [log in to unmask] containing the line
UNSUBSCRIBE SOCNET in the body of the message.