201301 Qr Mf Varda
Transcript of 201301 Qr Mf Varda
-
8/10/2019 201301 Qr Mf Varda
1/71
Coding Qualitative Data for Social
Network Analysis
Danielle M. Varda, PhDAssistant Professor, School of Public Affairs
Cameron Ward-HuntPhD Candidate, School of Public Affairs
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
2/71
Outline for Todays Talk
What is SNA?
How is social network data (typically)
collected?
How is social network data coded?
Using qualitative data for SNA
Two (maybe three) examples Issues with social network data
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
3/71
WHAT IS SNA?
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
4/71
Social Network Analysis
Social Network Analysis (SNA) is a tool used
to gather and analyze data to explain the
degree to which network actors connect toone another and the structural makeup of
collaborative relationships (Scott, 1991).
Allows new leverage for answering standard
social and behavioral science research
questions (Wasserman and Faust 1994)
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
5/71
Basic Assumptions of Network Analysis
Relationships Matter
People Influence Each Other
Ideas and materials flow through relationships
Structure of relationships have consequences
Not just composition of elements of system
that matters, but also how they are put
together (how they are embedded within a
system)
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
6/71
Elements of SNA
Collects data on who is connected to whom
How those connections vary and change
Focus patterns of relations Distinct from the methods of traditional
statistics and data analysistheories, models,
and applications are expressed in terms ofrelational concepts or processes.
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
7/71
What is a Network?
A set of nodes (or actors) along with a set of
ties of specified type that link them.
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
8/71
Elementsof a Network: Nodes
Set of actors (nodes) connected by a set of ties
Individuals
Organizations, departments, teams
These nodes have attributes
Any description of the node
Often characterized by
groups (e.g. gender, sector)
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
9/71
Elements of a Network: Ties
Ties connect pairs of actors
Directed (i.e., potentiallyone-directional, as in givingadvice to someone)
Undirected (as in beingphysically proximate)
Dichotomous (present orabsent, as in whether twopeople are friends or not)or
Valued (measured on ascale, as in strength offriendship)
2
2
1
3 3
3
2
1
1
1
1
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
10/71
Why Study Networks?
Stop the spread of disease How relationships influence our health
behaviors
The spread of innovative practices Study how organizations partner to leverage
resources
Anti-terrorism For quality improvementto improve
performance
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
11/71
Meaning in Nodes & Lines
SNA provides an additional way to evaluate relationships
Current Assumption = More is better.
More partners = successful collaboration (counting noses)
Alternative Assumption = Less can be more.
Not based on how many partners you have, but how they are
connected.
New
Relationship
YOU YOU
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
12/71
SNA is Informed by Theories
Diffusion
Contagion: Likelihood that network members willdevelop beliefs, assumptions, and attitudes that aresimilar to those of others in their network
Exchange and Dependency
Resource dependency
Homophily, Proximity, and Social Support Theories
Evolutionary & Coevolutionary Theories Ecological Approaches
Age, size dependence; technological processes, communityinterdependent; organizational change
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
13/71
2 Different Network Approaches
Whole Network
A complete set of bounded actors
Example: All members in a tobacco coalition, all public
health departments in the country, all clients in ahealth delivery network
Ego/Personal Network
Randomly sample people from a population
Ask only about their alters (no roster) Ask a sample of patients about who the members of
their personal support network are
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
14/71
Unit of Analysis: Whole/Sociocentric Level
NetworksVary in Size,
Shape, and
Composition
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
15/71
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
16/71
Krackhardts Kite Network - (Centrality)
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
17/71
Unit of Analysis: Subgroup Level
Subgroups
are a subsetof the graph
based on
certainnodes or
links
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
18/71
Unit of Analysis: Dyads/Triads
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
19/71
Unit of Analysis: Individual Nodes (Ego-Centric)
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
20/71
HOW IS SOCIAL NETWORK DATA(TYPICALLY) COLLECTED?
20
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
21/71
Data Collection &Management
1. Identify the population
Bounding the network, gaining access
2. Determine the data sources
Archival, interviews, observations, surveys
3. Collect the data
Instrument design
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
22/71
Identifying the Population: Bounding the Study
22
Extremely vexing to beginners and outsiders Network concept would seem to argue against boundaries
Empirical research makes clear we are all connected Even if distant links dont matter, some people in the sample will be atthe edge, no matter where we cut it
Identify a boundary
Theoretical
Affiliation (Members of; Friend of)
Defined Groups (Coalitions; Employees of an Organization; Children in a Classroom)
Stakeholders (not so clear?)
Pre-Data Collection Work Might Be Necessary
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
23/71
Step 2: Determine Data Sources
23
Archival data/Text Analysis
Covert Networks
Citation Networks
Meeting Minutes
Surveys (online, paper, interviews; can include
network questions as part of survey)
Observations
Data Mining (internet, emails)
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
24/71
Sampling??
Can you use a sampling method to studycomplete networks? In general, the answer isno.
Exception: Egocentric However, whole networks are not sampled
purpose is to survey the whole network!
There may be exceptions.
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
25/71
Step 3: Collect the Data
Surveys are typically either: Name Generator.
unlimited in scope: the respondent may name anyone from anysphere of life: neighbors, kin, friends, coworkers, etc.
After obtaining a large list of names, the interviewer typically goes overeach name, asking the respondent about the nature of their relationshipwith that person (what social relation) and asking about attributes of thatperson (sex, race, income, etc.).
Bounded List
Pre-defined list Entire network must be identified before data collection starts
Sometimes boundaries are clear (e.g. classrooms, organizationaldepartments)
Sometimes not clear; might need to implement name generatorapproach first
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
26/71
Survey Data Collection Methods
Questionnaires.
Row-based: each questionnaire forms one row in
the adjacency matrix of the group as a whole.
Use the whole matrix analytically
Each row obtained from a different source
Each could have its own measurement
idiosyncracies
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
27/71
Example Survey Questions
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
28/71
Example Survey Questions
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
29/71
Example Survey Questions
WHO: Name of other
organization or group
partnership?
Get specifics, e.g., dept
or unit, location,
contact name(s).
Also note name of the
partnership itself (if ithas one).
TIMING:How
long has the
partnership
been going?
Is it ongoing vs.
past work?
If ended, when
and why?
CONTENT: What kinds of activities does
the Partnership entail?
Mark all that apply from response to
question. Do not read each category
below, but may use them to prompt
respondent if having difficulty answering.
ROLES: Is
there a lead
agency or set
of agencies in
the
partnership?
RESOURCES: Is there
any dedicated funding for
the Partnership, either
within the partner
organizations or from
sources outside the
Partnership?
Focus on type of support(and sources for outside
support), but not on
amount of funding.
OUTCOME:
How successful
has it been and
why? (specific
to the individual
partnership
listed below)
# ___ a Years ___
b Months ___
1 Ongoing
2 Ceased
When & Why?
1 Conduct research 9 Tools
Develop
2 Conference 10 Training
3 Educational program 11 Tech
Assistance
4 Info Dissemination 12 Legal/RegulChange
5 Intellectual Exchang13 New
Technologies
6 Fund Research 14 Data Repositories
7 Standards Develop 15
Advocacy/Awareness
8 Guidelines Develop 16
Other: ___________
1 No
2 Yes :
____________
____________
____
1 Monetaryeither org
2 In-kind support only
(default)
3 Monetaryoutside
source
Source(s):
_____________________
____________________
1 Successful
2 Somewhat
successful
3 Not
successful
4 Too early to
tell
Notes:
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
30/71
Adding An Ethnographic Approach
30
Ethnography at front end helps to
Select the right questions to ask
Word the questions appropriately
Create enough trust to get the questions
answered
Ethnography at the back end helps to
Interpret the results
Can sometimes use resps as collaborators
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
31/71
HOW IS SOCIAL NETWORK DATACODED?
31
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
32/71
1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 21 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9
R R R R R R R R R R R R R R R A A A A A A A A A A A A A A
- - - - - - - - - - - - - - - - - - - - - - - - - - - - -
1 R1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 1 0 0 0 0 1 0
2 R2 0 1 1 1 0 1 1 1 1 1 1 1 0 1 1 0 0 0 0 0 0 0 0 0 1 0 0 0
3 R3 1 1 0 1 1 1 1 1 1 1 1 1 1 0 1 0 0 0 0 1 0 0 0 0 0 0 0 1
4 R4 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 0 0 0 0 0 0 0 0 1 1 0 0 0
5 R5 1 1 1 0 1 1 0 1 1 1 1 1 0 1 1 0 0 0 0 0 0 1 0 0 0 0 0 0
6 R6 0 1 1 1 0 1 1 0 1 1 1 0 1 1 1 1 0 0 0 0 1 0 0 0 0 0 1 0
7 R7 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 0 0 0 0 0 0 1 0 0 0 1 0 0
8 R8 1 1 0 1 1 0 1 0 1 1 0 1 1 0 1 0 0 0 0 0 0 0 1 0 0 0 0 1
9 R9 0 1 0 1 1 0 0 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0
10 R10 1 1 1 0 1 1 1 0 1 0 0 0 1 1 1 0 0 0 1 0 0 0 1 0 0 1 0 0
11 R11 1 1 1 1 1 1 1 0 0 0 1 1 1 1 1 0 0 0 0 1 1 0 0 0 0 0 0 0
12 R12 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0
13 R13 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 1 0 0 0 0 1 0
14 R14 1 1 1 0 0 1 1 1 1 1 1 1 1 1 1 0 0 1 0 0 0 0 0 0 0 0 0 0
15 R15 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0
16 A1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
17 A2 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 1 0 1 1 0 1 0 1 1 1 1 1
18 A3 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 1 1 1 0 0 1 1 1 1 1 1 019 A4 0 0 0 0 0 0 1 0 0 0 0 1 0 1 0 1 1 0 0 0 0 0 0 0 0 1 1 1
20 A5 1 1 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 1 1 1 1 1 1 0 1 1
21 A6 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 1 1 0 0 1 1 1 1 1 1 1 1 1
22 A7 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 1 1 1 1 1 1 1 0 1
23 A8 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 0 1 1 1 1 1 0 0 0
24 A9 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 1 1 1 1 1 1 0 1 0 1 0 1 0
25 A10 0 1 0 0 0 0 1 0 1 0 1 0 0 0 0 1 1 1 1 1 0 1 1 0 1 1 1 1
26 A11 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 1 1 0 1 1 0 0 0 0 1
27 A12 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 1 1 1 1 0 1 0 1 0 1 0 0
28 A13 0 1 1 1 0 0 0 0 0 1 0 0 1 0 0 1 1 1 1 1 0 1 1 1 1 1 1 0
29 A14 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 0 1
Data is Entered Into an Adjacency Matrix
Question: Who do
you work with?
A 1 indicates the presence
of a relationship.
A 0 represents theabsence of a relationship.
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
33/71
Network Logical Data Structures
Ed Sue Jim Bob
Ed - 1 0 0
Sue 0 - 1 1
Jim 0 0 - 0
Bob 1 0 0 -
Ed Sue Jim Bob
Ed - 4 0 2
Sue 0 - 5 1
Jim 0 0 - 0
Bob 3 0 4 -
Friendship
Email Communicat ion
Individual characteristics onlyhalf the story...RELATIONSMATTER!
People influence each other,ideas & material flow
Values are assigned to pairs ofactors
Hypotheses can be phrased interms of correlations betweenrelations
*2012 LINKS Center Summer SNA Workshop: Analyzing Track
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
34/71
Relational Data & Attribute Data
Ed Sue Jim Bob
Ed - 1 0 0
Sue 0 - 1 1
Jim 0 0 - 0
Bob 1 0 0 -
Gender Education Salary
Ed 0 14 50000
Sue 1 15 99000
Jim 0 12 65000
Bob 0 8 15000
Relational Data Attribute Data
SNA provides the ability to combine relational data withattribute data (e.g., homophily, heterogeneity, etc)
*2012 LINKS Center Summer SNA Workshop: Analyzing Track
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
35/71
Graphical representation of a digraph
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
36/71
USING QUALITATIVE DATA FOR SNA- 3 EXAMPLES
36
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
37/71
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
38/71
Codeword Barbarossa
Operation Barbarossa
Surprise German
invasion of the Soviet
Union in 1940
Primary Source: Codeword Barbarossa,
complied by historian Barton Whaley(1973)
Documents84 sub-cases with relevant
information exchanges 38
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
39/71
Excerpt
49. A Warning from TitoIn mid-May, while the German divisions in conqueredGreece and Yugoslavia were hurriedly being routed throughBelgrade toward Rumania, another opportunity for acredible disclosure existed. Vladmie Dedijer reveals in hisofficial biography of Tito: A senior German officer told aRussian refugee that Hitler was preparing to attack Russia.This information reached Tito, who sent a radiogram toDimitrov toward the end of May bringing it to his notice.Dimitrov, in Moscow in his capacity as secretary-general of
the Comintern, would have immediately informed theNKVD, if not other Soviet authorities, of such intelligence.
39
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
40/71
GeneralGeorgyZhukov
GeneralS.I.Kabanov
GeneralSikorski
GeneralissimoChiang
Kai-shek
GeorgiDimitrov
GermanSergeant-Ma
jorDeserter
GustavHilger
HansHeinrichHerwarthvonBittenfield
HansLazar
HaroldH.Tittleman
HarryCarlson
HarryFlannery
Josef Masin
Josef Stalin
Josef Tito 3
Khlopov
Konon Molody
Konstantain Umansky
Laurence Steinhardt
Leopold Trepper
Lieutenant Colonel Louis Baril
Lieutenant Commander Alwin (The Shadow) Kramer
Lieutenant-General Ivanovich Golikov
Lieutentant-General M.A. Purkayev 6
Lord Casey
Louis Lochner
Coding ExampleInfo Case Date Narrative Type Value
Press leak 46 15-May-41 From Hans Lazar to Kowalewski through Pangal - to Polish govt in exile Message 6
AP channel 47 23-May-41 From Beck to Maass to Lochner to TASS to GRU and NKGB
Leaked
Document 9
GRU in Berlin 48 22-May-41 From Khlopov to GRU headquarters Message 6
Tito 49 May-41 From Unk German Officer 2 to Unk russian refugee to Tito to Dimitrov to NKVD Message 3
Napoleonic Clue 53 1-Jun-41 From Etzdorf to Lanza Message 3
Map clue 51 May-41 Observable in Court photographer window to Berezhkov Observable 1
Counterfeit Rubles 52 Jun-41 Observable to Kelly Observable 2
40
Coding of the
Information Exchange
Network
1) Extraction
2) Matrix coding
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
41/71
LouisLochner
Khlopov
JosefTito
UNKGermanOfficer2
UNKRussianRefugee
GeorgiDimitrov
Dr.HassovonEtzdorf
MicheleLanza
AdmiralKuznetsov
AdmiralFrancoisDarlan
Louis Lochner
Khlopov
Josef Tito 24
UNK German Officer2 23
UNK Russian Refugee 23
Georgi Dimitrov
Dr. Hasso von Etzdorf 23
Michele Lanza
Admiral Kuznetsov
Admiral Francois Darlan
Coding Example
41
Name Position Nationality Location Echelon Position
Admiral Kuznetsov Commisar of the Navy Soviet Russia StrategicAndre de Vodianer Hungarian Minister Hungarian Portugal Diplomatic
Andrey Vyshinsky Soviet Deputy Foreign Commisar Soviet Russia Diplomatic
Carlo GRU agent Soviet France Tactical Covert
General Georgy Zhukov Chief of Staff, Moscow Soviet Russia Strategic
General S.I. Kabanov Commanding Officer, Soviet Base Hango Peninsula Soviet Sweden Strategic
Georgi Dimitrov Comintern Secretary General Bulgarian Russia Diplomatic
Ivan Filippov Chief of TASS Bureau-Berlin Soviet Germany Strategic Covert
Ivan Maisky Soviet Ambassador to the United Kingdom Soviet England Diplomatic
Josef Stalin General Secretary Soviet Russia Head of State
Khlopov Deputy Military Attache in Berlin Soviet Germany Strategic
Social Network Coding
For Attribute File
For Network Matrix
(also in binary)
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
42/71
Example FindingsRQ3: How do nations share intelligence information?
Figure 3
Barbarossa Social Network by Nationality
Soviet (N=34), German (N=31), American (N=26), British (N=12).
19 nationalities
18 locations
Strong international
social network =
Potential for
communication
But does the potential network
translate to information shared?
Soviet
German
American
42
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
43/71
Example FindingsRQ3: How do nations share intelligence information?
Figure 4
Barbarossa Information Network by Nationality
43.8% of all transactions
occurred between
participants of different
nationalities
Conclusion
Robust percentage of sharing
outside of diplomatic channels
Different sharing patterns (i.e.
Americans versus British)
13.6% shared by
diplomatic ties
American
Soviet
2%
12%
6%
12%
68%
Brokerage Roles,InfoNet:Nationality
Coordinator
Gatekeeper
Representative
Consultant
Liaison
43
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
44/71
EXAMPLE 2: COLLABORATING FOR
IMPACT: USING SOCIAL NETWORK
ANALYSIS TO EXPLORE NONPROFITCOMMUNITY INTERCONNECTIONS
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
45/71
Data
Dataset drawn from a community of nonprofitorganizations Online website, GivingFirst, where nonprofit organizations in the
greater Metro Denver area post detailed profiles of theirorganizations in order to raise funding for their organizations.
Databas Variables we coded included:
Number of staff in the organization (including full-time, part-time, volunteer, and contractors),
Governance information (number and names of the Board of
Directors members), Revenue information,
Mission or purpose of the nonprofit organization, and
Each organizations partnerships and affiliations
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
46/71
Example of Text We Coded
46
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
47/71
How We Coded This
47
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
48/71
Data
Respondent (organizations that posted profiles),N= 362
These 362 organizations identified 2219 otherorganizations as either partners or affiliates
In total, 3765 dyads (or relationships) weregenerated. Of these dyads, 3149 were identified by respondents
as collaborations and 616 as affiliations.
The data analysis was performed only on the 3149collaborations.
UCINET used for exploratory SNA
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
49/71
Connectivity
Fully Connected All nodes reachable
Most with 1 (N=1087), 2 (N=301), 3 (N=135), 4 (N=181), 5 (N=97), 6(N=30), 7 (N=84), 8 (N=36), 9 (N=87), 10 (N=131), 11 (N=54), 12(N=46), 13 (N=9),
Layers of connectivity
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
50/71
Components
One large component; 21 other components
Made up of policy areas: Behavioral Health, Courts/OffenderPrograms, Dance/Theater, Environmental, Faith-Based, Health,International Development, International Human Rights, CountyOrganizations, Music (Band), Parochial Schools, Prisons/Reentry,Rotary, Spanish Arts, Sports (Soccer), Young Adults, Water, some
uncategorized because orgs not consistently servicing one area. Not grouped by NTEE-CC categories
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
51/71
Key PlayersInDegree & OutDegree
Out egree
American Humane Association 119
Colorado Humanities 85
Share Our Strength's Operation Frontline CO 69
AfricAid, Inc. 65
Parenting Place 55
Autism Society of Colorado 50
Street's Hope 48Cross Community Coalition 41
ACCESS Housing 40
Colorado Dragon Boat Festival 33
In egree
Denver Public Schools 26
University of Denver 18
Food Bank of Rockies 14
Denver Health Medical Center 13
Mile High United Way 13
Head Start 12
Colorado Nonprofit Association 11
Colorado Coalition for the Homeless 10
SafeHouse Denver, Inc 9
Family Tree, Inc. 9
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
52/71
Brokerage
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
53/71
Discussion Points
Nonprofit Communities are highly connected Connections tend to form based policy areas,
rather than NTEE categorization Connections are based on need (resource
dependency; access to client population) etc. Connections within groups tend to be
Coordinating positions Certain types of categories act more as brokers than
others
Organizational capacity seems to have somethingto do with # of connections Betweeness seems to have more to do with the
description of the clients served
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
54/71
EXAMPLE 3:
COLLECTING DATA FROM A
COMMUNITY COALITION TO INFORM
QUALITY IMPROVEMENT
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
55/71
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
56/71
Who: Early Learning; Family
Support & Parent Education;
Social Emotional & Mental Health;
Health
Purpose: To identify stakeholders
and ideal system
Collecting DataA Hands On Approach
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
57/71
What the Groups Produce
57
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
58/71
Coding the Pictures of Ideal Systems
58
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
59/71
ISSUES WITH SNA DATA
59
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
60/71
Issues with SNA Data
60
Response bias
Asymmetry
Missing data
Accuracy
Ethics
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
61/71
Ethical Issues
61
Respondents cannot be anonymous
Non-respondents are still included
Missing data can be powerful
Has the potential to be mis-used by
Management
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
62/71
Data Collection Limitations
Informant accuracy Can people really tell you about their social networks?
Marketing researchers have found that consumers can barelytell you what they had for lunch yesterday. Bernard, Killworth
and Sailer investigated informant accuracy systematically andfound that about 52% of what they said was wrong.
Based on the work of Freeman, Freeman and Romney, as wellD'Andrade, DeSoto, and many others, it appears that people'srecall of their interactions with others is systematically biased
toward what is normal and/or logical.
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
63/71
Data Collection Limitations
People also tend to remember interactions withpeople who are important, while forgettinginteractions with people that are not.
Some respondents will lie to make themselves lookgood, since people judge others on who theyassociate with.
As with any questionnaire, there are also problemswith how people interpret the questions. What
"friend" means to one person may be very differentfrom what friend" means to others.
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
64/71
Resources
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
65/71
SNA Professional Organization
wwww.INSNA.org
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
66/71
Comprehensive List of Courses
http://socialnetworkcourses.wordpress.com/2010/11/11/list-of-snsna-courses/
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
67/71
Office of Behavioral & Social Sciences Research
http://obssr.od.nih.gov/scientific_areas/methodology/systems_science/index.aspx
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
68/71
List of Recommended Readings
http://obssr.od.nih.gov/pdf/valente_recomen_readings.pdf
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
69/71
UCINET
http://www.analytictech.com/ucinet/
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
70/71
Online SNA Text (UCINET)
http://www.faculty.ucr.edu/~hanneman/nettext/
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspxhttp://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx -
8/10/2019 201301 Qr Mf Varda
71/71
PARTNER
(Program to Analyze, Record, and TrackNetworks to Enhance Relationships)
www.partnertool.net
http://www.ucdenver.edu/academics/colleges/SPA/Pages/index.aspx