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

8/10/2019 201301 Qr Mf Varda

3/71

WHAT IS SNA?

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)

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)

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.

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.

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)

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

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

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

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

8/10/2019 201301 Qr Mf Varda

14/71

Unit of Analysis: Whole/Sociocentric Level

NetworksVary in Size,

Shape, and

Composition

8/10/2019 201301 Qr Mf Varda

15/71

8/10/2019 201301 Qr Mf Varda

16/71

Krackhardts Kite Network - (Centrality)

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

8/10/2019 201301 Qr Mf Varda

18/71

Unit of Analysis: Dyads/Triads

8/10/2019 201301 Qr Mf Varda

19/71

Unit of Analysis: Individual Nodes (Ego-Centric)

8/10/2019 201301 Qr Mf Varda

20/71

HOW IS SOCIAL NETWORK DATA(TYPICALLY) COLLECTED?

20

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

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

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)

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.

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

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

8/10/2019 201301 Qr Mf Varda

27/71

Example Survey Questions

8/10/2019 201301 Qr Mf Varda

28/71


8/10/2019 201301 Qr Mf Varda

29/71


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:

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

8/10/2019 201301 Qr Mf Varda

31/71

HOW IS SOCIAL NETWORK DATACODED?

31

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.

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

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

8/10/2019 201301 Qr Mf Varda

35/71

Graphical representation of a digraph

8/10/2019 201301 Qr Mf Varda

36/71

USING QUALITATIVE DATA FOR SNA- 3 EXAMPLES

36

8/10/2019 201301 Qr Mf Varda

37/71

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

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

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

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)

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

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

8/10/2019 201301 Qr Mf Varda

44/71

EXAMPLE 2: COLLABORATING FOR

IMPACT: USING SOCIAL NETWORK

ANALYSIS TO EXPLORE NONPROFITCOMMUNITY INTERCONNECTIONS

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

8/10/2019 201301 Qr Mf Varda

46/71

Example of Text We Coded

46

8/10/2019 201301 Qr Mf Varda

47/71

How We Coded This

47

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

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

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

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

8/10/2019 201301 Qr Mf Varda

52/71

Brokerage

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

8/10/2019 201301 Qr Mf Varda

54/71

EXAMPLE 3:

COLLECTING DATA FROM A

COMMUNITY COALITION TO INFORM

QUALITY IMPROVEMENT

8/10/2019 201301 Qr Mf Varda

55/71

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

8/10/2019 201301 Qr Mf Varda

57/71

What the Groups Produce

57

8/10/2019 201301 Qr Mf Varda

58/71

Coding the Pictures of Ideal Systems

58

8/10/2019 201301 Qr Mf Varda

59/71

ISSUES WITH SNA DATA

59

8/10/2019 201301 Qr Mf Varda

60/71

Issues with SNA Data

60

Response bias

Asymmetry

Missing data

Accuracy

Ethics

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

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.

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.

8/10/2019 201301 Qr Mf Varda

64/71

Resources

8/10/2019 201301 Qr Mf Varda

65/71

SNA Professional Organization

wwww.INSNA.org

8/10/2019 201301 Qr Mf Varda

66/71

Comprehensive List of Courses

http://socialnetworkcourses.wordpress.com/2010/11/11/list-of-snsna-courses/

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

8/10/2019 201301 Qr Mf Varda

68/71

List of Recommended Readings

http://obssr.od.nih.gov/pdf/valente_recomen_readings.pdf

8/10/2019 201301 Qr Mf Varda

69/71

UCINET

http://www.analytictech.com/ucinet/

8/10/2019 201301 Qr Mf Varda

70/71

Online SNA Text (UCINET)

http://www.faculty.ucr.edu/~hanneman/nettext/

8/10/2019 201301 Qr Mf Varda

71/71

PARTNER

(Program to Analyze, Record, and TrackNetworks to Enhance Relationships)

www.partnertool.net