Social Network Analysis

businessunknownInternet and Web Development

Nov 12, 2013 (3 years and 8 months ago)

142 views

Template Updated 1 May 2011

Social Network Analysis

Mark A. Sellers
,
30 July 2011

Mercyhurst College, Erie PA

Advanced Analytic Techniques Course


Purpose

The purpose of this project is to assess the value of
using the analytical method of Social
Network Analysis to determine the answer to the following analytical question:


H
ow are organized crime families organized to maintain cohesiveness and control
the flow of inform
ation, influence and resources?


For my
dataset, I will create a data table showing the personal and business relationships of
characters from Francis Ford Coppola’s film Godfather I. The criteria that I will use to evaluate
this method will include:




Ability to describe individual interpersonal

relationships accurately.



Utility of UCINET Social Network Analysis software.



Ability of different centrality measures (Degree,
Closeness,
Betweenness and
Eigenvector) to characterize relative importance and influence of individual members in
the network.



Utility of creating trust, task, money and resource, and strategy and goal networks in
understanding influence and information flow in organized crime families.


Literature Review


An overview of the topics addressed in the papers cited in the Detailed Literature Review in
Annex A shows that social network analysis is a technique that has very broad application in
both criminal and counter
-
terrorist intelligence: Van der Hulst, Schwa
rtz and Rouselle and
Natarajan (2006) focused their studies on understanding criminal organizations, while Brams et
al (2005), Koschade (2006) and Basu (2005) studied terrorist networks. Although there is a
common thread of applying SNA to different netwo
rks, the authors emphasize not only different
output measures of SNA in their respective papers, but several also focus on ancillary issues such
as gathering and preparing data and preparing the analysis for the customer.


The book by Hanneman and Riddle
(2005),
Introduction to Social Network Methods
, is a general
guide to the theoretical and practical aspects of SNA methodology and illustrates just how
extensive this analysis technique is. Dozens of output variables characterizing the importance of
indivi
dual actors, such as centrality and power indices; the macroscopic structures of networks,
such as network density; and the characteristics of sub
-
network structures, such as cliques and
factions are available in the output of light
-
weight desktop applicat
ions such as UCINET and
other analysis packages. Given that there are so many ways to characterize social networks, it is
perhaps not surprising that different authors use variations on the basic technique to characterize
the networks that they study.


The

discussion by van der Hulst (2009) of SNA as a tool for criminal investigations describes the
basic use of centrality and other output variables in characterizing criminal networks, but places
businessunknown_f8c7c425
-
112e
-
4cb1
-
9ddc
-
ef0b862a9ef3.docx


2

more emphasis on the use of these results for purposes such as

scenario building, hypothesis
testing, targeting, and preparation of evidence. The author also outlines issues in collecting and
preparing data for network analysis, such identifying and characterizing affiliations, collecting
demographic data and elimina
ting redundant data. This paper considers SNA as an end
-
to
-
end
process, from identifying a target group to final reporting.


The study of Schwartz and Rouselle (2009) on using SNA to target criminal networks is much
more theoretical than the other sources
cited. The authors propose new indices for assessing
Network Capital (NC) and Intelligence Worth (IW) as aides to targeting individuals for
enforcement or intelligence gathering. Individuals with high NC would be likely targets for
arrest, due to their abi
lities to make decisions and influence events, while those with high IW
would be likely targets for surveillance. Although the results of this study are promising, neither
NC nor IW is available in off
-
the
-
shelf applications such as UCINET.


Brams et al. (
2005) discuss the creation of directed graphs (digraphs) for creating social network
diagrams to show the influence paths among actors within those networks. The methodology
they propose must be applied manually; however, their modifier is easily applied t
o small
networks to show their hierarchical structure.


Koschade’s (2006) analysis of the Jemaah Islamiyah terrorist network responsible for the 2002
Bali bombing includes a straightforward SNA of the terrorist cell using more commonly used
indices such as

network size, density, and degree, closeness and betweenness centrality. Like
van der Hulst (2009), Koschade discusses preparation of the data for analysis and like Natarajan
(2006) he stresses the value of using other analytical techniques such as a tho
rough case study to
support the analysis.


Unlike the other papers, which focus on individual actors in social networks, Basu (2005) applies
SNA to the study of networks of organizations. Although the actual analysis presented includes
only the betweenness

measures for organization in a network, her study underscores the value of
applying SNA to organizations of networks as well as individuals.


Natarajan (2006) discusses the use of SNA as part of a criminal investigation for the preparation
of a prosecutor
’s case against a heroin trafficking network. Like van der Hulst (2009) and
Koschade (2006), Natarajan does discuss her SNA is some detail, but also describes the
exigencies of data collection and the value of using SNA as part of a suite of analytical
te
chniques.


From the small sampling of sources on SNA described above, it is obvious that SNA may be
defined broadly or narrowly. Those who describe it broadly consider SNA as an end
-
to
-
end
process from collection to reporting, while those who consider it n
arrowly focus strictly on the
actual “number
-
crunching” that is done by desktop applications. As in many other studies not
cited here, at least one of the studies cited (Brams et al. (2005)) uses a previously described data
set (Krebs (2002)) as surrogate
data for their technique.


Each of the studies cited in this small survey of available references addresses a unique aspect of
SNA. SNA is a new field and is still growing. New modifiers continue to be added and new
businessunknown_f8c7c425
-
112e
-
4cb1
-
9ddc
-
ef0b862a9ef3.docx


3

analysis of criminal and terrorist netwo
rks add to our understanding of the advantages and
pitfalls of using SNA as an intelligence tool.


Description

Social Network Analysis
(SNA)
is a
methodology

for identifying and characterizing
relations

between individuals
,

factions

or other subgroups
within social organization
al structures
.
These
organizations, or
social network
s,

can be
found in
formal
ly established
organizations
, such as
corporation
s
, trade union
s
, or professional societ
ies
,
or in ad hoc groups or coalitions assembled
for a single p
urpose

for a short period of time
, as well as in criminal or terrorist networks and
organizations.


There are two basic components of any social network: nodes and links. Nodes are
generally

the
individuals
within the network,
but
nodes can
also
be used t
o depict factions, teams, or other
subgroups (of course these types of nodes would have their own
internal
social network
structure). Links are the ties or relationships between the
nodes

and these may be of several
different types.
These would include
t
rust

relationships, such as blood relationships or
friendships;
task

relationships, such as employer
-
employee or contract relationships;
money and
resource

relationships, such as
those with bankers or clandestine financiers; and
strategy and
goal

relations
hips, such as those between key leaders and their trusted lieutenants. These
relationships may be quantified using a predetermined scale such as 1 = weak relationship, 5 =
strong relationship).


SNA uses quantitative measures that are derived from a field of mathematics known as
graph
theory
. Consequently, S
NA
should

be done
using

a computer. Fortunately, there are several
SNA programs available (some are freeware), perhaps the most popular of which is UCINET,
which is produced by Analytic Technologies and which is used in this project.


The importance of individual nodes may be c
alculated by calculating their
centrality

in the
network. Centrality can be of three types
:

degree centrality, betweenness centrality and
closeness centrality. Degree centrality is a measure of the total number of links between a node
and the other nodes

in the network. Between
n
ess centrality measures the
importance of a node
in acting as a bridge between groups of nodes within a network. Closeness centrality quantifies
the number of links that connect a node to the other nodes.

These are illustrated i
n Figure 1.


SNA is useful in understanding the structures of criminal and terrorist organizations and can be
used in a number of ways, including scenario building, risk analysis and threat assessment,
hypothesis testing, identification of structural weakn
esses for exploitation, identification of
redundant actors and redundant roles, decision support for employing intelligence assets and
preparation of evidence for prosecution . It has been used in law enforcement and
counterterrorism analysis, as well as
for analysis of transnational issues such as narcotics
trafficking and weapons proliferation.


businessunknown_f8c7c425
-
112e
-
4cb1
-
9ddc
-
ef0b862a9ef3.docx


4


Figure 1. This figure illustrates different types of centrality. Node A connects two distinct subgroups in
the network, so it has high betweenness centralit
y. Node B directly connects to five other nodes, so its
degree centrality would be fairly high. Node C has fewer direct connections to other nodes than B, but its
paths into the red nodes are shorter, so it would have higher closeness centrality.

Node D w
ould have low
centralities for all three.


SNA is effective in identifying and understanding patterns of communication, authority,
influence and financial workings of organizations and informal networks that may not be
apparent from isolated bits of
information about individuals within the network. It can help
analysts identify not only leaders, but also brokers and funding sources within networks. This
methodology is a bit labor
-
intensive and dependent on accurate and complete information.


Strengt
hs



Identification of Key Players, Brokers, Bridge Spanners and other actors

Social Network Analysis identifies important actors within social networks by calculating
quantitative measures of their importance in different roles and functions
.

This provides
an advantage over link analysis, which is more qualitative in its approach.



Targeting

Social Network Analysis can provide indices of fragmentation and the reach of
individuals within networks. This can assist decision makers in determining w
hich actors
should be taken out of the network to cause the greatest degree of fragmentation and
which should be targeted for intelligence collection to provide the greatest amount of
information.



Common Operational/Intelligence
Picture

Network diagrams ca
n
summarize a tremendous amount of data into
a
n easily understood

picture
that can be shared between agencies
and
presented to decision makers.



Automated Proce
ssing

Once an accurate database is compiled, output can be derived easily using PC software.



We
aknesses



Requires Accurate and Complete Database


Algorithms cannot distinguish between good and bad data.
To calculate accurate
measures of importance and influence of network actors, data
entered

into the network
businessunknown_f8c7c425
-
112e
-
4cb1
-
9ddc
-
ef0b862a9ef3.docx


5

diagram or matrix must be complete,
accurate and free of redundancy
.

Identifying
missing data can be difficult.



Requires Large Investment Of Time And Effort

Actors in criminal and terrorist networks
work hard to keep their associations clandestine.
A significant degree of surveillance is likely to be required to discover and characterize
connections between actors.
Once gathered, d
ata must be carefully examined and vetted
for accuracy and completene
ss and entered into large and complex network diagrams.



Data Interpretation
C
an
B
e
S
ubjective

Determining the boundaries of networks can be difficult and subject to the interpretation
of the analyst. Different roles and relationships must be interpreted a
nd relationships can
change over time.


How
-
To
Apply Social Network Analysis

Social network analysis will allow the analyst to understand
the structure and inner relations of a
targeted group or organization, such as a terrorist network or organized crime
family
.

SNA can
also be an excellent collaborative tool as well as tool for briefing decision
-
makers interested in
targeting key actors for intelligence gathering or enforcement actions.


The preliminary work of identifying the members of a network and gat
hering accurate and
complete information about each member can be
quite
time
-
consuming. The quality of the
analysis, however, depends on gathering as complete a picture as possible. Groups that operate
clandestinely will do all that they can to conceal t
heir activities and
hide the inter
-
relationships
among members that allow them to do business. Therefore, the analyst should not expect to
complete the analysis quickly. In fact, social network databases will constantly evolve as new
members join the org
anization, as new alliances are formed with other organizations and as
members leave the organization.


The actual analytical phase of
SNA is quantitative and is best done on a computer
.

The analyst
should therefore be familiar with at least the basic features of desktop application such as
UCINET before attempting an analysis. Fortunately, such programs are readily available and
come with adequate documentation that explains each outpu
t variable in the analysis.


The basic steps of SNA include:


1.

Identify the Target Organization or Group


The first step is to identify the group that you are targeting
.

Although organized crime
and terrorist orga
nizations are nebulous, identifying a few k
ey actors associated with a
major crime of terrorist event is sufficient to get
you

started. The analytical picture is
almost certain to grow over time, but it is not necessary to have identified all of the
perpetrators before beginning the analysis
.

2.

Decid
e the Network Type that Best Characterizes the Targeted Organization and Fits the
Analytical Question

Criminal organizations can be complex, with different func
tions compartmentalized to
control information. For example,
t
he actors involved in money laund
ering, drug
trafficking, prostitution and other profit
-
making enterprises may be quite separate from
those involved in extortion or murder for hire. When possible, it can be helpful to focus
businessunknown_f8c7c425
-
112e
-
4cb1
-
9ddc
-
ef0b862a9ef3.docx


6

on trust, task, money/resource and strategy/goal sub
-
networks

to
keep the analysis
manageable
.

3.

Gather Information on Relationships between Key Actors in the Organization


It is important to gather as much information as possible to understand who knows
whom, who trusts whom, and who supervises or tasks whom as well as w
ho is trusted
with information, who is skilled in performing specific tasks or functions and who is
trusted to manage money or resources
.

4.

Vet the Information
, Eliminate Redundant or Inaccurate Information and Identify
Intelligence Gaps

Careful review of all data is
essential

to an accurate analysis.
R
elationships between
actors
must be clearly understood

and actors within the network must be accurately
identified
. Redundancies in data can be misleading and skew the results of the analysi
s;
for example, if a certain member of the organization is known by his given name and an
alias, it is possible that he may appear as two separate individuals.
As the network
diagram evolves, it may also be possible to identify missing members and direct
gathering efforts to fill these intelligence gaps.

5.

Create the Social Network Diagram

As data
are

gathered, the network diagram should be constructed. There are two possible
ways of doing this in UCINET software. The first is to construct the diagram directly in
an application called NETDRAW. This is not the
recommended

option as diagrams can
get comp
lex and details can be lost. The superior alternative is to construct the diagram
by constructing an adjacency matrix or ordinal matrix. An adjacency matrix simply
shows which member has a relationship with which members by assigning a value of 0
for no re
lationship and 1 for any type of relationship. An example is shown in
Table

1
.


Table

1
. Example of an adjacency matrix.


Al

Bob

Charlie

Don

Ed

Al

-

1

0

1

0

Bob

1

-

1

0

1

Charlie

0

1

-

0

0

Don

1

0

0

-

1

Ed

0

1

0

1

-


Little
can be known from this
type of matrix beyond the fact that relationships exist
. For
example, the matrix
shows
that Bob and Al know each other

t
here is a
1

in the Bob row
in the Al column and a 1 in the Al row in the Bob column

b
ut not much more than that.


An ordinal matrix, o
n the other hand, provides information on the strength of
relationships and the direction of influence in relationships. An example of this is shown
in
Table

2
.


Table 2
. Example of an ordinal matrix.


Al

Bob

Charlie

Don

Ed

Al

-

5

0

0

0

Bob

5

-

4

0

3

Charlie

0

2

-

0

0

Don

3

0

0

-

1

Ed

0

3

0

1

-


businessunknown_f8c7c425
-
112e
-
4cb1
-
9ddc
-
ef0b862a9ef3.docx


7

In this example, we can see that Bob and Al have a close relationship (5) and that neither
exerts influence over the other. Don, however, strongly influences Al (as seen by the 3 in
the Don row of the Al
column), but Al has little influence over Don (as seen by the 0 in
the Al row of the Don column).


The UCINET software will automatically construct a social network diagram from an
adjacency or ordinal matrix, making this task much easier than painstaking
creating the
diagram by hand.


6.

Use the Software to Compute Centrality and other Measures for Members of the Networks

Obtaining the centrality measures for degree, betweenness and closeness is a simple
menu
-
driven process. In addition to these basic measures, UCINET will produce a
number of other indices that characterize both the individual members and the network as
a
whole. These measures will be printed in tabular form and can be saved for reporting.
These measures, in addition to careful examination of the network diagram itself, may be
useful targeting
network
members for enforcement actions or intelligence gather
ing.


Steps in producing a social network analysis are shown as a concept map in Figure
2
.


businessunknown_f8c7c425
-
112e
-
4cb1
-
9ddc
-
ef0b862a9ef3.docx


8


Figure
2
. Concept map showing fundamental steps in social network analysis.


For Further Information

The following sources are excellent sources for understandi
ng the basic concepts of social
network analysis (Van der Hulst, Heuer and Pherson, and Weaver) and the application of the
UCINET software (Borgatti et al). A number of useful and insightful examples and case studies
may be found at the ORGNET web site.


Van der Hulst, R. (2009). Introduction to Social Network Analysis (SNA) as an investigative
tool.
Trends in Organized Crime
, 12(2), 101
-
121.


Heuer, R. J. & Pherson, R. H. (2011).
Structured Analytical Techniques for Intelligence
Analysis
.

Washington, DC: CQ Press.


Weaver, L. E. (2006). Social Network Analysis. In

K. J. Wheaton, E. E. Mosco & D. E. Chido
(Eds.),

The Analyst’s Cookbook, Volume 1
. (pages 103

116). Erie, PA: Mercyhurst College
Institute of Intelligence Press.

businessunknown_f8c7c425
-
112e
-
4cb1
-
9ddc
-
ef0b862a9ef3.docx


9


Borgatti, S.

P., Everett, M. G., & Freeman, L. C. (2002).
UCINET 6 for Windows. Software for
Social Network Analysis. User's Guide
. Analytic Technologies, Harvard Massachusetts, USA.
49 pp. Retrieved from: http://www.analytictech.com/ucinet/help.htm


Orgnet:
http://www.orgnet.com/index.html
.

Much of the pioneering work of Valdis Krebs and
his colleagues can be found on this site


Personal Test Case


Analytic Question and Evaluation Criteria

This study is intended to ans
wer the analytic question:


How are organized crime families organized to maintain cohesiveness and control
the flow of information, influence and resources?


According to the following criteria:




Ability to describe individual interpersonal relationships
accurately.



Utility of UCINET Social Network Analysis software.



Ability of different centrality measures (Degree,
Closeness,
Betweenness and
Eigenvector) to characterize relative importance and influence of individual members in
the network.



Utility of cre
ating trust, task, money and resource, and strategy and goal networks in
understanding influence and information flow in organized crime families.


Steps

Identif
ication of
the Target Organization

In this study, the target organization group used is actuall
y a surrogate data set. The target
organization is the fictional Corleone Crime family as it existed from 1946 to 1955 as depicted in
Francis Ford Coppola’s classic film,
The Godfather
.


Network Type
Based on
Analytical Question

(
How are organized crime fa
milies organized to
maintain cohesiveness and control the flow of information, influence and resources?
):

The network type that
best fits the analytical question is a corporate or hierarchical organized
crime family. This type of network is not an
ad hoc

or loosely amalgamated
communal business
network, such as a modern terrorist network. The Corleone family is based on traditional Italian
organized crime structures such as the
Mafia
(Sicily),
Camorra

(Campania)
,

'Ndrangheta

(Calabria) or
Sacra Corona Uni
ta

(Pulia). Since membership in such organizations is a formal,
life
-
long commitment, such organizations are permanent structure with well
-
defined boundaries.



Relationships between Key Actors in the Organization

Surrogate d
ata

were gathered from repeated viewings of the DVD version of
The Godfather
,
with each scene replayed as often as necessary to understand and characterize the relationships
among the members of the network. Each link or relationship was characterized as trus
t, task,
money/resource or strategy/goal. In most cases, each link could be characterized by more than
one of these categories.


businessunknown_f8c7c425
-
112e
-
4cb1
-
9ddc
-
ef0b862a9ef3.docx


10

Quality Control
/
Vet
ting
and Preparation
of
Information

The primary source for initially establishing all links was the DVD of t
he film. Entries for each
scene were subsequently cross
-
checked against the content of all or most of the other references.
These references are listed in Table 3. which also includes an index of the reliability of each
source, estimated in accordance wit
h
http://daxrnorman.googlepages.com/WebSitesYouCanTrust
-
Oct08.pdf



In constructing the adjacency matrices, the strength of each connection between characters is
based on the
numbers of times that any two characters appear in the same scene together. This
allows links between characters to be scored or quantified, e. g., if two characters appear together
in three different scenes, their link score is three. This approach is b
ased on the assumption that
emphasis on familial relations (father
-
son, brothers, etc.) will be shown in more scenes than
friendships, business partnerships, and other non
-
kinship relations.


In Table 2, each relation was characterized as Trust, Task, Mone
y/Resource and/or
Strategy/Goals. In most cases, relationships do not fall into a single category. Family members
engaged in a common enterprise will share trust relations, money/resource relations, and
strategy/goal relations. All trust, money/resource
, and strategy/goal relations are assumed to be
bi
-
directional (depicted as

), while task relations are unidirectional (

). For task relations,
the symbol


may be interpreted as “works for” or “is tasked by”. For example, Vito
Corleone

Luca Brasi may
be interpreted as Luca Brasi works for Vito Corleone.

In
each

relation, the senior or
superior

member of the pair is shown first. In the example above, Vito
Corleone is superior to Luca
Brasi
.


Table
3
. Sources used in this study. The primary source for

each line item in Table 2 is the original
film, The Godfather. Additional sources were used to confirm or clarify the observed relationships.

SOURCE

RELIABILITY

Coppola, F. F. (Director). (1972). The Godfather. Los Angeles: Paramount Pictures.

High
(60.09)

Bruno, A. [date unknown]. Fact and Fiction in the Godfather. TrueTV Crime Library

Criminal Minds and Methods. Retrieved July 22, 2011 from:
http:/
/www.trutv.com/library/crime/gangsters_outlaws/mob_bosses/the_godfather/1.html


Moderate

(35.18)

The Godfather Wiki. [date unknown]. Retrieved July 22, 2011 from:
http://godfather.wikia.co
m/wiki/The_Godfather_Wiki


Moderate

(35.18)

The Godfather

Official Site. [date unknown]. Retrieved July 22, 2011 from:
http://www.thegodfather.com/


High (56.41)

The Godfather Trilogy: The Godfather (1972). [d
ate unknown]. Retrieved July 22,
2011 from:
http://www.thegodfathertrilogy.com/


High (49.30)

The Official Mario Puzo Library

The Godfather. [date unknown]. Retrieved July 22,
2011 from:
http://www.mariopuzo.com/godfather/godfather.shtml


High (56.41)

Puzo, M. and Coppola, F. F. (1971). The Godfather Screenplay. New York:
Paramount Pictures. Retrieved July 22, 2011 from:
http://www.dailyscript.com/scripts/The_Godfather.html


High (49.30)


Social Network Diagrams

To create the
social network diagrams shown below, data from the table in Annex 2 was used to
create ordinal matrices for
t
rust,

t
ask,
m
oney/
r
esource and
s
trategy
/g
oal networks. In all cases the
social network diagrams are directional graphs (digraphs). For the
trust,
money/resource and
strategy/goal

diagrams all linkages
follow the relations denoted as

bi
-
directional (

), while for
the t
ask network diagram, the linkages are uni
-
directional (

).


businessunknown_f8c7c425
-
112e
-
4cb1
-
9ddc
-
ef0b862a9ef3.docx


11

Analytic Question Data

Data for this study is included in a
table of linkages in Annex 2.
This data was used to create the
matrices used by UCINET to create network diagrams and calculate the centrality measures
discussed below.


Results and Discussion


Trust Network

The Trust network is shown in Figure
3
. This ne
twork includes all members of the Corleone
family and their associates who have some kinship of friendship tie, but not necessarily (or
exclusive of) any business ties. Node size is proportional to degree centrality, while the
thickness of the connections

is proportional to the number of co
-
occurrences of the characters in
scenes in the film. These measures will be described below.


From the basic diagram, it is obvious that the main trust bonds are within extended family
members and life
-
long friends (and

one mistress). The Corleones are, of course, related by
blood, while Tom Hagen is an adopted son of Vito Corleone. Kay Adams, Carlo Rizzi, Theresa
Hagen, Apollonia Vitelli, and Sandra Corleone are related by marriage. The caporegimes, Peter
Clemenza an
d Sal Tessio are life
-
long friends and business partners of Vito Corleone, while
Paulie Gatto has a personal friendship with Fredo Corleone. Johnny Fontaine is a god
-
son of
Vito, while Don Tommasino, who guards Michael in Sicily is a friend of Vito. Lucy

Mancini is
Santino Corleone’s mistress.


The smaller trust network includes Philip and Bruno Tattaglia (father and son, respectively) and
Virgil Sollozzo.



Figure
3
. The trust network of the Corleone family and associates

created in UCINET
.


businessunknown_f8c7c425
-
112e
-
4cb1
-
9ddc
-
ef0b862a9ef3.docx


12

Because
members of trust networks may or may not be engaged in business dealings, the trust
network is not as likely as other networks to explain or reveal the centralities of network
members engaged in illegal activities.


Task Network

The Task network is show in

Figure
4
. In this diagram, members and links are shown without
regard to centrality or tie strength for the sake of clarity.



Figure
4
. The
task

network of the Corleone family and associates

created in UCINET
.


Following methodology described in
Brams et al. (2005), several Mutual Influence Sets (MIS)
may be identified by inspection of the diagram. At the highest level, Michael, Santino and Vito
Corleone form an MIS that tasks different levels, but is not tasked. The consigliore, Tom Hagen,
and t
he capo régimes, Clemenza and Tessio, form a second MIS, which is immediately
subordinate to the Corleones. A third MIS includes soldiers who may be tasked by the
consigliore/capo régimes, which includes Rocco Lampone, Paulie Gatto and Willi Cicci. More
t
rusted soldiers (Al Nieri and Luca Brasi) are tasked directly by the highest MIS, but have no
tasking authority. Also included are several network members who are not involved in criminal
activities (Bonasera, Nazorine, Enzo the Baker), who are also direc
tly tasked by the highest level
of the MIS.


Particularly interesting is the tasking relationship between Moe Greene and Fredo Corleone (who
is not tasked by his father or his brothers). This clearly shows Fredo’s duplicitous role in the
family effort to
take over the gambling industry in Las Vegas.


By capturing the traditional hierarchical structure of a Sicilian crime family, the methodology of
Brams et al. (2005) illustrates that SNA is capable of preserving not only the personal ties that
solidify and

organization, but the hierarchical ones as well.


businessunknown_f8c7c425
-
112e
-
4cb1
-
9ddc
-
ef0b862a9ef3.docx


13

Money/Resource Network

The Task network is show in Figure
5
. In this diagram, the strengths of ties is not shown for the
sake of clarity; however, an analysis of factions within the network was performed
in UCINET,
from which the colors of the nodes represents the six factions found in the network.



Figure
5
. The
Money and Resource

network of the Corleone family and associates

created in
UCINET
.


The diagram in Figure
5

clearly shows the Corleone fami
ly (including the capo régimes) as a
single faction. Also included in this faction is Don Tommasino, a friend of Vito’s, who takes
care of Michael while he is in Sicily. The “pentagram” in the lower left hand side of the diagram
clearly shows the “commis
sion”, which includes the heads of four of the five families plus Vito
Corleone (oddly enough, no fifth family head is ever identified in the film). The episode in which
Vito Corleone makes movie producer Jack Woltz “an offer he can’t refuse” to get a role

for
Johnny Fontaine in a new war movie shows up as a separate faction in the analysis, as does the
collusion between Fredo and Moe Greene when Michael takes over Greene’s interests. Although
these events are separated by nearly ten years in the course of

the film, in the network diagram
there is no such temporal separation. It is noteworthy that the analysis identified as these as
separate events by identifying the actors as separate factions.


The faction analysis is not perfect, however. The relation
between McCluskey and Sollozzo
appears as a single faction; however, Bruno Tattaglia is not identified as part of this faction.
This is most likely due to his tie to his father, Philip, which divides his loyalties between his
business partner, Sollozzo, a
nd his father. Luca Brasi appears to be in a faction with Bruno
Tattaglia, but like Tattaglia, his “loyalties” (Brasi feigned being a traitor to the Corleones to gain
the confidence of Sollozzo) are divided between Sollozzo and Vito Corleone.


Strategy/Goal Network

businessunknown_f8c7c425
-
112e
-
4cb1
-
9ddc
-
ef0b862a9ef3.docx


14

The Task network is show in Figure
6
. In this diagram, tie strength is proportional to line
thickness and the size of nodes is proportional to betweenness centrality.


This diagram clearly shows the Corleone family and the commission
as two distinct
organizations with separate goals and strategies. As a member of both organizations, Vito
Corleone has a very high degree of betweenness centrality, acting as a broker between the two
organizations.


Figure
6
. The Strategy and Goal netw
ork of the Corleone family and associates created in UCINET.


Of course, Virgil Sollozzo is not a part of the Corleone family, but has interacted with each
member in various capacities (Vito refuses to support Sollozzo’s heroin trade, Tom Hagen acts
as an

intermediary at the start of a gang war, and Michael and Santino plot to kill him).
Although Sollozzo has worked (unsuccessfully ) with the Corleones on this matter, his ties are
not nearly as strong as those between Vito and Michael, Santino and Tom Hag
en or as strong as
the ties between these three.


Although Bruno Tattaglia has connections to both sub
-
structures, his influence (based on tie
strength) is not significant, nor is his ability to act as a broker (low betweenness centrality).


Centrality and

other Measures for Members of the Networks

Centrality measures are calculated for each of the four network types described above (Trust,
Task, Money/Resource and Strategy/Goal). These centrality measures will include Degree,
Closeness, Betweenness and Eigenvector centralities. Degree centrality is
the simplest measure
of centrality and is the determined by counting the number of links connected to a member of the
network. Degree centrality does not necessarily give the best measure of the power of an
individual in a network, as key leaders may choo
se to insulate themselves by keeping their
contacts to a minimum and working through intermediaries. Closeness centrality is a measure of
the distance between a member and other members in the network. This distance is measured by
the path, or number of ot
her members through which a node must go to connect to another node.
businessunknown_f8c7c425
-
112e
-
4cb1
-
9ddc
-
ef0b862a9ef3.docx


15

Members who do not have to go through many nodes (i.e., who have the most direct contact),
have higher closeness centralities than those members who have to go through many nodes to
cont
act a particular node. Betweenness centrality is a measure of an individual’s position or
ability to act as an intermediary between cliques or factions within a network. Finally,
Eigenvector centrality is a measure of an individual’s power or influence w
ithin a network.


Trust Network
Centralit
ies

Centrality measures for the Corleone family Trust network is shown in Table
4
.


Table
4
. Centrality Measures for Trust Network


Degree

Closeness

Betweenness

Eigenvector

Vito

Corleone

25.714

5.477

1.58

53.63

Michael

Corleone

40

5.521

11.776

62.237

Santino

Corleone

28.571

5.486

6.202

53.46

Tom

Hagen

22.857

5.469

3.697

45.414

Fredo

Corleone

20

5.46

3.389

40.924

Peter

Clemenza

14.286

5.435

0

37.234

Sal

Tessio

14.286

5.435

0

37.234

Emilio

Barzini

0

0

0


Philip

Tattaglia

2.857

2.939

0

0

Carmine

Cuneo

0

0

0


Victor

Strachi

0

0

0


Virgil

Sollozzo

2.857

2.939

0

0

Bruno

Tattaglia

5.714

2.941

0.168

0

Luca

Brasi

0

0

0


Johnny

Fontaine

5.714

5.385

0

17.121

Al

Nieri

0

0

0


Willi

Cicci

0

0

0


Connie

Corleone

14.286

5.435

0.067

35.992

Carlo

Rizzi

14.286

5.426

0.28

33.32

Mama

Corleone

8.571

5.401

0.067

20.167

Fabrizio


0

0

0


Don

Tommassino

2.857

5.376

0

9.197

Apollonia

Vitelli

5.714

5.385

0

10.791

Signore

Vitelli

5.714

5.385

0

10.791

Amerigo

Bonasera

0

0

0



Nazorine

0

0

0


Enzo

the Baker


0

0

0

Lucy

Mancini

2.857

5.344

0

7.9

Theresa

Hagen

2.857

5.327

0

6.711

Kay

Adams

5.714

5.393

0

15.907

Sandra

Corleone

2.857

5.344

0

7.9

Jack

Woltz

0

0

0


businessunknown_f8c7c425
-
112e
-
4cb1
-
9ddc
-
ef0b862a9ef3.docx


16

Table
4
. Centrality Measures for Trust Network

Paulie

Gatto

2.857

5.319

0

6.047

Captain

McCluskey

0

0

0


Moe

Greene

0

0

0


Rocco

Lampone

0

0

0



The trust family network is based on kinship and friendship ties rather than business ties. Even
though relationship are not based on business of financial interests, it is not surprising that
Michael Corleone scores highest on all centrality measures, followed closely by his father Vito.
The key members of this network have the greatest number of connections (degree centrality),
the fewest number of intermediaries (closeness centrality), the g
reatest ability to broker deals and
decisions (betweenness centrality) and the most power (Eigenvector centrality).


Task Network
Centralit
ies

Task network centralities are shown in Table
5
. Of note in this table is that Vito and Michael
Corleone and Tom H
agen have the greatest number of connections (degree centrality), but that
their closeness centralities are not significantly higher than those of the capo régimes (Clemenza
and Tessio) or even minor characters like Luca Brasi or Johnny Fontaine. This is
a much smaller
network, with much closer ties among the members. Betweenness centrality is a clear indicator
of who is able to task and who is tasked (betweenness centrality = 0). Eigenvector centrality
clearly shows who is powerful and who is not.


T
able
5
. Centrality Measures for Task Network


Degree

Closeness

Betweenness

Eigenvector

Vito

Corleone

20

4.692

5.546

50.079

Michael

Corleone

22.857

4.704

7.395

55.864

Santino

Corleone

8.571

4.648

0.084

35.463

Tom

Hagen

22.857

4.717

5.07

67.566

Fredo

Corleone

2.857

2.857

0

0

Peter

Clemenza

17.143

4.704

3.641

54.934

Sal

Tessio

11.429

4.685

0.952

44.402

Luca

Brasi

2.857

4.605

0

10.641

Johnny

Fontaine

5.714

4.66

0.504

22.51

Al

Nieri

2.857

4.617

0

11.87

Willi

Cicci

2.857

4.63

0

14.356

Fabrizio


2.857

4.617

0

11.87

Amerigo

Bonasera

5.714

4.654

0

24.997


Nazorine

2.857

4.605

0

10.641

Enzo

the Baker

2.857

4.617

0

11.87

Paulie

Gatto

2.857

4.617

0

11.672

Moe

Greene

2.857

2.857

0

0

Rocco

Lampone

8.571

4.673

0

37.898


Money/Resources Network
Centralities

businessunknown_f8c7c425
-
112e
-
4cb1
-
9ddc
-
ef0b862a9ef3.docx


17

Centrality measures for the Money/Resources network are shown in Table
6
. In this network,
degree centralities reflect the numbers of connections between members. Closeness measures do
not vary much among the members, which is not unexpected
in a comparatively small network.
Betweenness measures clearly show that Vito and Michael Corleone are the main brokers in
financial deals, although Virgil Sollozzo ranks high due to his unsuccessful attempt to broker an
arrangement with the Corleones and

the Tattaglias to manage the heroin trade in New York City.
Finally, Eigenvector centralities capture the relative power of the members of the network. Vito,
Michael and Santino Corleone and Tom Hagen have the highest Eigenvector centralities, with
Clem
enza, Tessio and Sollozzo also scoring above 40. The heads of the other families (Barzini,
Philip Tattaglia, Cuneo and Strachi) also rank fairly high, but are not in the same league as the
Corleones. Barzini, who is seen more often associating with Vito
Corleone, has the highest
Eigenvector centrality of the family heads.


Table
6
. Centrality Measures for Money/Resources Network



Degree

Closeness

Betweenness

Eigenvector

Vito

Corleone

31.429

5.486

8.531

62.612

Michael

Corleone

28.571

5.486

11.172

52.54

Santino

Corleone

17.143

5.443

0.136

46.697

Tom

Hagen

20

5.452

2.433

47.988

Fredo

Corleone

2.857

5.224

0

1.395

Peter

Clemenza

14.286

5.426

0

40.962

Sal

Tessio

17.143

5.435

0.812

44.867

Emilio

Barzini

14.286

5.385

0.276

28.187

Philip

Tattaglia

14.286

5.385

0.636

23.846

Carmine

Cuneo

11.429

5.376

0

21.971

Victor

Strachi

11.429

5.376

0

21.971

Virgil

Sollozzo

20

5.452

3.261

41.392

Bruno

Tattaglia

8.571

5.36

0.224

13.532

Luca

Brasi

8.571

5.376

0.108

18.902

Johnny

Fontaine

5.714

5.36

0.392

9.948

Don

Tommassino

2.857

5.344

0

8.45

Jack

Woltz

5.714

5.327

0.084

9.317

Captain

McCluskey

5.714

5.368

0

15.106

Moe

Greene

5.714

5.36

2.857

8.674


Strategy/Goal Network Centralities

For the Strategy/Goals Network (Table
7
), the most conspicuous result is the
high betweenness
centrality for Vito Corleone, which clearly shows him as the single link between the Corleone
organization and the commission. Vito Corleone’s Eigenvector centrality shows him to be much
more powerful than even his sons and his consiglior
e, as well as the heads of the other families.


Table
7
. Centrality Measures for Strategy/Goals Network


Degree

Closeness

Betweenness

Eigenvector

Vito

Corleone

22.857

3.7

2.605

71.313

businessunknown_f8c7c425
-
112e
-
4cb1
-
9ddc
-
ef0b862a9ef3.docx


18

Michael

Corleone

11.429

3.684

0

41.822

Santino

Corleone

11.429

3.684

0

41.822

Tom

Hagen

11.429

3.684

0

41.822

Emilio

Barzini

11.429

3.684

0

41.822

Philip

Tattaglia

14.286

3.688

0.588

45.082

Carmine

Cuneo

11.429

3.684

0

41.822

Victor

Strachi

11.429

3.684

0

41.822

Virgil

Sollozzo

14.286

3.688

0.588

45.082

Bruno

Tattaglia

5.714

3.676

0.084

18.85



Conclusions

A number of conclusions may be drawn from this study in answering the analytic question posed
at the beginning of this paper, namely, “
How are organized crime families organized to maintain
cohesiveness and
control the flow of information, influence and resources?
” Following
Natarajan’s (2006) typology of corporate organizations versus communal business organizations,
it is clear that the Corleone family is a corporate organization. Connections based on kin
ship and
friendship ties are fundamental to holding the organization together, but a hierarchical structure
exists to ensure that the organization functions in its purpose. A clear leader (Godfather) heads
the organization, and is strongly supported by his

sons. Capo régimes answer to the Godfather,
and soldiers in turn answer to them. A consigliore supports the Godfather with counsel and by
acting as an intermediary in negotiating disputes with other criminal organizations. This is
organizational structu
re and function is well
-
known to those with even cursory familiarity with
depictions of organized crime families in popular fiction or true
-
crime books.


The more immediate question is h
ow well does
social network analysis
accomplish the goals of
th
is

proj
ect
, i.e., how well does the methodology capture the structure and function of an
organized crime family? The short answer to this is, very well indeed!


A visual inspection of the social network diagrams will clearly show who is powerful, who is
well
-
conn
ected and who is not. In the trust network diagrams, Michael Corleone emerges as the
center of the Corleone organization, as seen in the number of ties he has with other members
(reflected in the size of the node in the diagram representing Michael, which
is proportional to
his degree centrality). When the method of Brams et al. (2005) is applied to the task network
diagram, Mutual Influence Sets are apparent which reflect the hierarchical structure of the
family. Social ties in the trust network reflect
the emotional connections between members, but
a definite corporate structure to the organization is shown in the task network.


In the money/resource network diagram, factions are quite apparent. These reflect temporary
arrangements created for specific b
usiness deals, including forcing Jack Woltz to hire Johnny
Fontaine, the Sollozzo heroin deal, and the arrangement between Fredo and Moe Greene that is
counter to Michael Corleone’s bid to take over Las Vegas. The role of Vito Corleone as the only
broker b
etween the Corleone family and the Commission could not be more obvious in the
strategy/goal network diagram.


businessunknown_f8c7c425
-
112e
-
4cb1
-
9ddc
-
ef0b862a9ef3.docx


19

These are qualitative measures and depictions of the structure and function of a corporate crime
family. The quantitative measures of SNA also ac
curately capture and describe the structure. In
the trust network centralities, Vito Corleone and his sons the bedrock of the family structure, as
shown by their degree centralities (most ties), closeness centralities (fewest intermediaries),
betweenness
centralities (position as brokers) and Eigenvector centralities (power). Likewise, all
centrality measures for the task network show who is and who is not powerful, but also indicate
who is at the top of the tasking chain and who is at the bottom. This c
onforms entirely with the
task network diagram that shows a clear hierarchy. In the money/resources centralities, we can
clearly identify who is more powerful, and in the strategy/goal centralities, Vito Corleone’s
position between his family and the Comm
ission can be seen just as it is in the network diagram.


Overall, social network analysis explains and quantifies the structure of a corporate criminal
organization, the relative power and influence of its members, and the roles they are likely to
play in

tasking subordinates and brokering business deals thoroughly

and unambiguously.

In
summary:




SNA is able to reflect
individual interpersonal relationships accurately
; however,
subjective evaluation is necessary in characterizing links as trust, task, mon
ey/resource
and/or strategy/goal before inputting data into the analysis
.



UCINET Social Network Analysis software

is quite capable of generating clear social
network diagrams and quickly calculating centrality measures that characterize the
network accurat
ely
.

UCINET is easy to use and capable of much more detailed
parameterization than that included in this study.



C
entrality measures (Degree,
Closeness,
Betweenness and Eigenvector)
accurately
characterize relative importance and influence of individual me
mbers in the network.



T
rust, task, money and resource,

and strategy and goal networks greatly aid
in
understanding influence and information flow in organized crime families.

Modifiers
such as that of Brams et al. (2005) add value to this analysis.




businessunknown_f8c7c425
-
112e
-
4cb1
-
9ddc
-
ef0b862a9ef3.docx


20

Annex 1. Detailed Literature Review

The following are critiques of various sources that discuss
social network analysis
.


Introduction to Social Network Analysis (SNA) as an investigative tool

Van der Hulst, R. C. (2009). Introduction to Social Network Ana
lysis (SNA) as an investigative
tool.
Trends in Organized Crime
, 12(2), 101
-
121.


Technique: Social Network Analysis


Purpose

The purpose of this article is to inform the reader about applying Social Network Analysis
(SNA) as an investigative tool in
understanding criminal and terrorist networks (adversary
networks), as well as to propose a draft protocol on applying SNA methodology from the
preparation phase to the reporting phase.


Strengths and Weaknesses


Strengths:



SNA p
rovides an objective method

for characterizing adversary networks (better than
link analysis



Quantifies relationships between actors in adversary networks (providing a more detailed
picture than that link analysis)



Versatile and broadly applicable. SNA can be used for a number of pu
rposes in law
enforcement and intelligence, including:

o

Scenario building

o

Risk analysis and threat assessments

o

Hypothesis testing

o

Identification of structural weaknesses for exploitation

o

Identification of redundant actors and redundant roles

o

Decision suppor
t for employing intelligence assets

o

Preparation of evidence for prosecution


Weaknesses:



Access to and collection of network data can be very difficult



Definition of network boundaries is not straightforward.



Different types of relations must be defined an
d behaviors must be interpreted



Data are not always reliable and sources are not always trustworthy



Relationships change over time. Bonds can become stronger or weaker



Gathering and evaluating data is time intensive



Requires software for data processing,
especially with large data sets



Data must be “tidied” for redundancy. Same actor can show up under several names



Missing data is not easy to spot and can lead to erroneous results


Applying the Method

Van der Hulst proposes the following protocol for appl
ying SNA:

Preparation:



Define category of the target group (e.g. criminal, terrorist) and network boundaries

businessunknown_f8c7c425
-
112e
-
4cb1
-
9ddc
-
ef0b862a9ef3.docx


21



Formulate research questions



Identify analysis routines to answer questions



Formulate assumptions



Develop coding system for attributes, activities
and affiliations

Data Processing:



Gather information (time intensive)



Identify attributes activities and affiliations of actors



Create database of attributes (age, sex, skills, criminal record)



Create incident matrix of affiliations



Sort the names of acto
rs



Tidy up data

Analysis and Reporting:



Consider routines and robust measures



Perform analysis routines



Interpret results



Report


Van der Hulst recommends UCINET as a useful software program for SNA and also includes a
tentative classification scheme for
network roles bases on a protocol devised by the Royal
Canadian Mounted Police.


Uses of SNA

Van der Hulst focuses on the use of SNA in understanding and characterizing criminal networks,
but cites a number of studies of the application of this method in

understanding terrorist
networks.


Most informative aspects of the paper

This is source is an excellent overview of the subject and provides and end
-
to
-
end description of

how to apply SNA as an investigative tool in understanding adversary networks. Although the
article does not go into detail on interpreting results of output from SNA tools or into the best
ways to present the results of SNA, the discussion of gathering,

sorting, editing and tidying data
for analysis is invaluable.


About the Source
Author

Renée C. van der Hulst is the director of the
Office

of
Network Analysis, an independent
consultancy firm in the Netherlands and
was employed by the Ministry of Justice

of the
Netherlands when she authored this paper. She is a member of the International Association for
the Study of Organized Crime and the International Network for Social Network Analysis. She
is the author of numerous scholarly papers on SNA.


Source r
eliability rating:

This source received a rating of 60.09: Very High Credibility.


Contact:

Mark A. Sellers,
mark.sellers@yahoo.com
,
Mercyhurst College, Erie PA, Advanced Analytic

Techniques Course, 19 June 2011.

businessunknown_f8c7c425
-
112e
-
4cb1
-
9ddc
-
ef0b862a9ef3.docx


22


O
ther sources cited
:

Blau PM (1964)
Exchange and power in social life
.
New York:
John Wiley & Sons

Borgatti SP, Jones C, Everett MG (1998) Network measures of social capital.
Connections

21(2):27

36

Borgatti SP, Everett M
G, Freeman LC (2002)
Ucinet 6 for Windows: Software for social network
analysis
. Analytic Technol, Harvard

Borgatti SP, Carley KM, Krackhardt D (2006) On the robustness of centrality measures under
conditions of imperfect data.
Soc Networks

28:124

136

Bour
dieu P (1986) The forms of social capital. In: Richardson JG (ed)
The handbook of theory:
Research for the sociology of education
. Greenwood Press, NY

Bruinsma GJN, Bernasco W (2004) Criminal groups and transnational markets: A more detailed

examination on

the basis of social network theory.
Crime Law & Soc Chang

41:79

94

Burt RS (1992)
Structural holes: The social structure of competition
. Harvard Univ Press,
Cambridge, MA

Burt RS (2000) The network structure of social capital. In: Sutton RI, Staw BM (eds)

Res in
organ behavior
. JAI Press, Greenwich, CT

Burt RS (2001) Structural holes versus network closure capital. In: Lin N, Cook KS, Burt RS
(eds)
Social capital: theory and research
. Aldine de Gruyter, NY

Carley KM (2003a) Dynamic network analysis. In: Br
eiger R, Carley K, Pattison P (eds)
Dynamic social network modeling and analysis: workshop summary and papers
. The Natl Acad
Press, Washington DC

Carley KM (2003b) Linking capabilities to needs. In: Breiger R, Carley K, Pattison P (eds)
Dynamic social
network modeling and analysis: Workshop summary and papers
. The Natl Acad
Press, Washington DC

Carley KM, Lee JS, Krackhardt D (2001) Destabilizing networks. Connections 24(3):79

92

Chattoe E, Hamill H (2005) It’s not who you know
-

it’s what you know abou
t people you don’t
know that counts.
Br J of Criminol

45:860

876

Clark RM (2007) Intelligence analysis: A target
-
centric approach. CQ Press, Washington, DC

Coleman JS (1990)
Foundations of social theory
. Harvard Univ Press, Cambridge, MA

Coles N (2001) It’
s not what you know
-

it’s who you know that counts: Analysing serious crime
groups as social networks.
Br J of Criminol

41:580

594

Contractor NS, Monge PR (2003) Using multi
-
theoretical multi
-
level (MTML) models to study

adversarial networks. In: Breiger
R, Carley K, Pattison P (eds)
Dynamic social network modeling
and analysis: Workshop summary and papers
. The Natl Acad Press, Washington DC

Cook KS, Whitmeyer JM (1992) Two approaches to social structure: Exchange theory and
network analysis.
Annu Rev of S
ociol

18:109

127

Emirbayer M, Goodwin J (1994) Network analysis, culture, and the problem of agency.
Am J of
Sociol
99 (6):1411

1454

Fijnaut C, Bovenkerk F, Bruinsma G et al (1998)
Organized crime in the Netherlands
. Kluwer
Law International, The Hague, th
e Netherlands

Flap HD, Völker B (2003) Creation and returns of social capital. Routledge, London Granovetter
M (1973) The strength of weak ties.
Am J of Sociol

78:1360

1380

Harary F (1969)
Graph theory
. Addison
-
Wesley, Reading, MA

Harper WR, Harris DH (1975) The application of link analysis to police intelligence.
Human
Factors

17 (2):157

164

Heider F (1958)
The psychology of interpersonal relation
. John Wiley & Sons, NY

businessunknown_f8c7c425
-
112e
-
4cb1
-
9ddc
-
ef0b862a9ef3.docx


23

Homans GC (1958) Social behavior as exchange.
Am J of Sociol

6
3:597

606

Kadushin C (2002) The motivational foundation of social networks.
Soc Networks

24(1):77

91

Kleemans ER, Brienen MEI, Van de Bunt HG (2002)
Georganiseerde criminaliteit in Nederland

[Organized crime in the Netherlands]. Boom Distributiecentrum, Me
ppel, the Netherlands

Klerks P (2001) The network paradigm applied to criminal organisations: Theoretical nitpicking
or a relevant doctrine for investigators? Recent developments in the Netherlands.
Connections

24(3):53

65

Koschade S (2006) A social networ
k analysis of Jemaah Islamiyah: The applications to
counterterrorism and intelligence.
Stud in Confl & Terror

29:589

605

Krackhardt D (1998)
S
immelian tie: Super strong and sticky In: Kramer RM, Neale M (eds),
Power and influence in organizations
. Sage, Th
ousand Oaks, CA

Krebs V (2002) Mapping networks of terrorist cells.
Connections

24(3):43

52

Leenders RThAJ, Gabbay SM (1999)
Corporate social capital and liability
. Kluwer Academic
Publishers, Boston, MA

Lin N (1992) Social resources theory. In Borgatta EF
, Borgatta ML (eds):
Encyclopedia of
Sociol
. NY, vol. 4:1936

1942

Lin N (2001)
Social capital: A theory of social structure and action
. Cambridge Univ Press,
Cambridge, NY

McDowell D (1998)
Strategic intelligence: A handbook for practitioners, managers and

users
.
Istana Enterprises Pty. Ltd, Cooma Australia

McGrath C, Krackhardt D, Blythe J (2003) Visualizing complexity in networks: Seeing both the
forest and the trees.
Connections

25(1):37

47

McGrath C, Blythe J, Krackhardt D (1997) The effect of spatial a
rrangement on judgments and
errors in interpreting graphs.
Soc Networks

19(3):223

242

McIllwain JS (1999) Organized crime: A social network approach.
Crime Law & Soc Chang

32:301

323

McNally D, Alston J (2006) Use of social network analysis (SNA) in the examination of an
outlaw motorcycle gang.
J of Gang Res

13(3):1

25

McPherson M, Smith
-
Lovin L, Cook JM (2001) Birds of a feather: Homophily in social
networks.
Annu Rev of Sociol

27:415

444

Mintzberg H (1983)
Structure in fives: Designing effective organizations, 6th translated edn
.
Prentice Hall, Englewood Cliffs, NJ, USA

Morselli C, Giguère C (2006) Legitimate strengths in criminal networks.
Crime Law & Soc
Chang

45 (3):185

200

Morsell
i C, Petit K (2007) Law enforcement disruption of a drug importation network.
Global
Crime

8 (2):109

130

Morselli C, Roy J (2008) Brokerage qualifications in ringing operations.
Criminol 4
6(1):71

98

Morselli C, Giguère C, Petit K (2007) The efficiency/secu
rity trade
-
off in criminal networks.
Soc
Networks

29:143

153

Natarajan M (2000) Understanding the structure of a drug trafficking organization: A
conversational analysis.
Crime Prevention Stud

11:273

298

Natarajan M (2006) Understanding the structure of a
large heroin distribution network: A
quantitative analysis of qualitative data.
J of Quant Criminol

22:171

192

Portes A (1998) Social capital: Its origins and applications in modern sociology.
Annu Rev of
Sociol

24 (1):1

24

businessunknown_f8c7c425
-
112e
-
4cb1
-
9ddc
-
ef0b862a9ef3.docx


24

Rothenberg R (2002) From whole c
loth: Making up the terrorist network.
Connections

24(3):36

Sageman M (2004)
Understanding terror networks
. Univ of Pennsylvania Press, Philadelphia

Schwartz DM, Rouselle DA (2008) Targeting criminal networks: Using social network analysis
to develop enfor
cement and intelligence priorities.
IALEIA J

18(1):18

44

Scott J (2000)
Social network analysis: A handbook
. Sage, Newbury Park CA

Simmel G (1950)
The sociology of Georg Simmel
. Free Press, NY

Sparrow MK (1991a) Network vulnerabilities and strategic intell
igence in law enforcement.
J of
Intell and Counterintell

5(3):255

274

Sparrow MK (1991b) The application of network analysis to criminal intelligence: An
assessment of the prospects.
Soc Networks

13:251

274

Strang S (2005)
User guide RCMP target profile sh
eet
. Royal Canadian Mounted Police, Canada
Tsvetovat M, Carley KM (2007) On effectiveness of wiretap programs in mapping social
networks.
J of Comput and Math Organ Theory

13(1):63

87

Tyler JR, Wilkinson DM, Huberman BA (2003) Email as spectroscopy: Automa
ted discovery of
community structure within organizations.
Communities and Technol

81

96

Van der Hulst RC (2004) Gender differences in workplace authority: An empirical study on
social networks. Univ Groningen (ICS thesis), Groningen, the Netherlands

Wasse
rman S, Faust K (1994)
Social network analysis: Methods and applications
. Cambridge
Univ Press, Cambridge

Williams (2001) Transnational criminal networks. In: Arquilla J, Ronfeldt D (eds)
Networks and
netwars: The future of terror, crime and militancy
. Ran
d Corporation, Santa Monica

Xu J, Marshall B, Kaza S et al (2004)
Intelligence security and security informatics
. Springer,
Berlin/Heidelberg


Introduction
t
o Social Network Methods

Hanneman, R. A. and Riddle, M. (2005).
Introduction to social network
methods
. Riverside, CA:

University of California, Riverside. Published in digital form at:
http://faculty.ucr.edu/~hanneman/




Technique: Social Network Analysis


Purpose

The purpose of this book is to present a basic introduction to the fundamental concepts of Social
Network Analysis (SNA) and the implementation of these concepts in practical analysis.
Although the authors state that this book is not a user’s manual for U
CINET software, the book
relies heavily on examples illustrated using UCINET. It would be difficult to envision using this
software without having this book open.


Strengths and Weaknesses



Strengths:



Comprehensiveness. Twelve of the eighteen
chapters are dedicated to interpretation of
SNA output, and each of these includes discussions of several output parameters. The
sheer number of output parameters produced in the analysis reveals just how
comprehensive SNA is as an analysis methodology.



T
he ability to quantify characteristics of individual actors within networks, including the
use of centrality measures to assess the influence if individual actors.



The ability to characterized the structure of social networks as a whole.

businessunknown_f8c7c425
-
112e
-
4cb1
-
9ddc
-
ef0b862a9ef3.docx


25



The ability to iso
late and identify cliques and sub
-
networks within broader social
networks.



The ability to visualize social networks fairly easily using NETDRAW software within
UCINET.


Weaknesses:



The book does explicitly identify any specific weaknesses of SNA.



Given the

scope of what SNA is able to produce, an apparent weakness of the method is
the steep learning curve associated with this method.


Applying the Method

This textbook describes how to determine and interpret every output variable that UCINET is
capable of
calculating, as well as the methodology of inputting SNA data into diagrams and
matrices. Describing the application of each feature of UCINET is beyond the scope of this
review.


Uses of SNA

This is a general textbook on applying SNA theory to virtually

any social network, ranging from
Cub Scout packs and local garden clubs to multi
-
national corporations, trade unions or terrorist
networks. The book focuses on applying theory to the practice, mainly using UCINET software
,
without focusing on any specific

application
.


Compar
ison with
other sources.

This is a far more basic reference than

R. C. Van der Hulst
’s


Introduction to Social Network
Analysis

(SNA) as an investigative tool”, which describes how one might approach the use of
SNA in intelligence anal
ysis or criminal investigation. Hanneman and Riddle’s book is a core
textbook for anyone using SNA in any context, particularly anyone using UCINET.


Most informative aspects of the paper

I would consider this text to be a core document for anyone doing
SNA
.

It provides perhaps the
most comprehensive practical overview of how to apply social network analysis, but it is far
more than a “cook book” or user’s manual for UCINET. The descriptions of how different
output parameters are calculated and what they

mean are clear and logical, without going into
mathematical detail.


About the Source
Author
s

Robert Hanneman is
a professor in the Department of Sociology in the College of Humanities,
Arts, and Social Sciences at the University of California, Riverside.

Mark Riddle is an assistant
professor in the Department of Sociology of the University of Northern Colorado.


Source reliability rating:

This source received a rating of 55.57: Very High Credibility.


Contact:

Mark A. Sellers,
mark.sellers@yahoo.com
,
Mercyhurst College, Erie PA, Advanced Analytic

Techniques Course,
25

June 2011.


businessunknown_f8c7c425
-
112e
-
4cb1
-
9ddc
-
ef0b862a9ef3.docx


26

O
ther sources cited
:

Alba,
Richard

D. 1973. "A graph
-
theoretic definition of a sociometric clique"
Journal of
Mathematical Sociology
, 3: 113
-
126

Alba,
Richard

D. and Gwen Moore. 1983. "Elite social circles" Chapter 12 in Burt and Minor
(eds).
Applied network analysis: A methodological introduction
. Beverly Hills: Sage.

Anheier
, Helmut K. 1987. "Structural analysis and strategic resea
rch design: Studying politicized
interorganizational networks"
Sociological Forum

2: 563
-
582

Arabie, P., S.A. Boorman, and P.R. Levitt. 1978. "Constructing blockmodels: How and why"
Journal of Mathematical Psychology
, Vol. 17: 21
-
63

Baker, Wayne E. 1986.

"Three
-
dimensional blockmodels"
Journal of Mathematical
Sociology
,
12: 191
-
223.

Baker
, Wayne E. 1990. "Market Networks and Corporate Behavior"
American Journal of
Sociology
. 96: 589
-
625.

Barnes
, J.A. 1983. "Graph theory in network analysis"
Social Netwo
rks

5: 235
-
244.

Berkowitz
, S.D. 1988. "Markets and market
-
areas: Some preliminary formulations" pp. 261
-
303
in Wellman and Berkowitz (eds.)
Social structures: A network approach
. Cambridge: Cambridge
University Press

Berkowitz
, S.D. 1988. Afterword: Towa
rd a formal structural sociology, pp. 477
-
497 in
Wellman and Berkowitz (eds.)
Social structures: A network approach
. Cambridge: Cambridge
University Press.

Berkowitz
, S.D. 1982.
An introduction to structural analysis: The network approach to social
resear
ch

Toronto: Butterworths

Berkowitz, S.D., Peter J. Carrington, Yehuda Kotowitz, and Leonard Waverman. 1978
-
79. The
determination of enterprise groupings through combined ownership and
directoship

ties,
Social
Networks

1: 391
-
413.

Blau, Peter and Robert K
. Merton. 1981.
Continuities in structural inquiry

London and Beverly
Hills: Sage.

Blau, Peter. 1981. Introduction: Diverse views of social structure and their common
denominator. in Blau and merton (eds.)
Continuities in structural inquiry
. London and Beverly
Hills: Sage.

Bodemann
, Y. Michal. 1988. Relations of production and class rule: The hidden basis of patron
-
clientage. pp. 198
-
220 in Wellman and Berkowitz (eds.)
Social structrues: A network approach
.
Cambridge: Cambridge University P
ress

Bonacich, Phillip. 1972. Technique for analyzing overlapping memberships. in Herbert Costner
(ed.)
Sociological Methodology 1972
. San Francisco: Jossey
-
Bass

Bonacich
, Phillip. 1972. Factoring and weighting approaches to status scores and clique
detection,
Journal of Mathematical Sociology
, 2: 113
-
120

Boorman
, S.A. 1975. A combinatorial optimization model for transmission of job information
through contact networks,
B
ell Journal of Economics
, 6: 216
-
249

Borgatti, Steven, Martin Everett, and Linton Freeman. 1992.
UCINET IV Version 1.0 User's
Guide
. Columbia, SC: Analytic Technologies.

Bott,
Elizabeth
. 1957.
Family and social network

London: Tavistock Publications

Bou
gon, Michel, Karl Weick, and Din Binkhorst. 1977. Cognition in organizations: An anlaysis
of the Utrecht jazz orchestra,
Administrative Science Quarterly
, Vol. 22: 606
-
639

Boyd,
John

Paul, John H. Haehl and Lee D. Sailer. 1972. Kinship systems and inverse

semigroups,
Journal of Mathematical Sociology
, 2: 37
-
61

businessunknown_f8c7c425
-
112e
-
4cb1
-
9ddc
-
ef0b862a9ef3.docx


27

Breiger, Ronald. 1979. Toward an operational theory of community elite structure,
Quality and
Quantity
, 13: 21
-
47.

Breiger
, Ronald L. and Philippa E. Pattison. 1978. The joint role structure of two

communities'
elites,
Sociological Methods and Research
, 7: 213
-
26.

Breiger
, Ronald L. and Philippa E. Pattison. 1986. Cumulated social roles: The duality of
persons and their algebras,
Social Networks
, 8: 215
-
256.

Breiger
, R.L. 1976. Career attributes a
nd network structure: A blockmodel study of a bio
-
medical research specialty,
American Sociological Review
, 41: 117
-
135

Breiger
, Ronald L. 1988. The duality of persons and groups. pp. 83
-
98 in Wellman and
Berkowitz (eds.)
Social structures: A network appr
oach
. Cambridge: Cambridge University
Press.

Brym, Robert J. 1988. Structural location and ideological divergence: Jewish Marxist
intellectuals in turn
-
of
-
the
-
century Russia. pp. 332
-
358 in Wellman and Berkowitz (eds.)
Social
structures: A network approac
h
. Cambridge: Cambridge University Press.

Burkhardt
, Marlene E. and Daniel J. Brass. 1989. Changing patterns or patterns of change: A
longitudinal investigation of the interplay of technology, structure, and power,
Pennsylvania
State University
, mimeo.

B
urt, Ro
n
ald S. 1982.
Toward a Structural Theory of Action
. New York: Academic Press.

Burt,
Ronald

S. 1983. Distinguishing relational contents. pp. 35
-
74 in Burt and Minor (eds.)
Applied network analysis: A methodological introduction
. Beverly Hills: Sage

Burt,
Ronald

S. 1983. Network data from informant interviews. Chapter 7 in Burt and Minor
(eds.)
Applied network analysis: A methodological introduction


Burt,
Ronald

S. 1983. Cohesion versus structural equivalence as a basis for network subgroups.
Chapte
r 13 in Burt and Minor (eds.)
Applied network analysis: A methodological introduction
.
Beverly Hills: Sage.

Burt,
Ronald

and M.J. Minor (eds.) 1983.
Applied network analysis: A methodological
introduction

Beverly Hills: Sage

Burt,
Ronald

S. 1982.
Toward
a structural theory of action: Network models of social structure,
perception, and action
. New York: Academic Press.

Burt,
Ronald

S. 1983.
Corporate profits and cooptation: networks of market constraints and
directorate ties in the American economy
. New Y
ork: Academic Press.

Burt,
Ronald

S. 1980. Models of network structure,
Annual Review of Sociology

6: 79
-
141

Burt, R
o
nald S. 1983. Network data from archival records. Chapter 8 in Burt and Minor (eds.)
Applied network analysis: A methodological introduct
ion
. Beverly HIlls: Sage

Burt, Ronald S. 1983. Range, Chapter 9 in Burt and Minor (eds.)
Applied network analysis: A
methodological introduction


Burt
, Ronald S. 1976. Position in networks,
Social Forces
, 55: 93
-
122.

Burt
, Ronald S. 1987. Social Contagion and Innovation.
American Journal of Sociology
. 92:
1287
-
1335.

Burt, Ronald S. 1992.
Structural Holes: The Social Structure of Competition
. Cambridge, MA:
Harvard University Press.

Carrington
, Peter J. and Greg H. Heil.
1981. COBLOC: A heierarchical method for blocking
network data,
Journal of Mathematical Sociology
, 8: 103
-
131

Carrington, Peter J., Greg H. Heil, and Stephen D. Berkowitz. 1979
-
80. A goodness
-
of
-
fit
index

for blockmodels,
Social Networks
, 2: 219
-
234

businessunknown_f8c7c425
-
112e
-
4cb1
-
9ddc
-
ef0b862a9ef3.docx


28

Cart
wright
, Dorwin. 1979. Balance and clusterability: An overview, pp. 25
-
50 in P. Holland and
S. Leinhardt (eds.)
Perspectives on social network research
. New York: Academic Press.

Cartwright
, Dorwin and Frank Harary.

1977. A graph theoretic approach to the investigation of
system
-
environment relationships,
Journal of Mathematical Sociology
, Vol. 5: 87
-
111.

Cartwright
, D. and F. Harary. 1956. Structural balance: A generalization of Heider's theory,
Psychological Revie
w
, 63: 277
-
92.

Clawson, Dan and Alan Neustadtl. 1989. Interlocks, PACs, and Corporate Conservatism.
American

Journal of Sociology
. 94: 749
-
73.

Clawson, Dan, Alan Neustadtl, and James Bearden. 1986. The Logic of Business Unity:
Corporate Contributions to
the 1980 Congressional Elections.
American Sociological Review
. 51:
797
-
811.

Coleman
, James, Elihu Katz, and Herbert Menzel. 1957. The diffusion of an innovation among
physicians,
Sociometry
, 20: 253
-
270

Cook,
Karen
. 1982. Network structures from an exch
ange perspective, in Peter Marsden and Nan
Lin (eds.).
Socal structre and network analysis
. Beverly Hills: Sage

Crane, Diana. 1969. Social structure in a group of scientists: A test of the 'Invisible college'
hypothesis
,
American Sociological Review
, 34:
335
-
352

Davis, J. 1963. Structural balance, mechanical solidarity, and interpersonal relations,
American
Journal of Sociology
, 68: 444
-
62

Davis, James A. 1967. Clustering and structural balance in graphs,
Human Relations
, 30: 181
-