tools for network interventions - inspq · tools for network interventions november 28, 2012...

Tools for Network Interventions

November 28, 2012

M t é l Q bMontréal, Quebec

Thomas W. Valente, PhDProfessor, Institute for Prevention Research

d k h l f dPreventive Medicine, Keck School of MedicineUniversity of Southern California

[email protected]

Simulation w/ 3 Initial Adopter Conditions

40

60

80

100

ent

Ad

op

ters

Opinion Leaders

Random

Marginals

0

20

1 2 3 4 5 6 7 8 9 10

Time

Per

ce

Marginals

16es Journées annuelles de santé publique 1

dursim01

Texte tapé à la machine

Cette présentation a été effectuée le 28 novembre 2012, au cours de la journée « Les liens sociaux et la contagion sociale : des vecteurs de santé » dans le cadre des 16es Journées annuelles de santé publique (JASP 2012). L’ensemble des présentations est disponible sur le site Web des JASP à la section Archives au : http//jasp.inspq.qc.ca/.

dursim01


dursim01


Steps to Opinion Leader InterventionsSteps to Opinion Leader Interventions

1) Collect/obtain network data)

2) Identify opinion leaders

3) Recruit them as champions

4) Convert them (if need be)

5) A i t th i th i b h i5) Assist them in their behavior change promotions

3

There are Several Issues

1. Other Centrality Measures

2. Stage of Diffusion

3. Bridges

4. Marginals

5. Low Threshold


Network PositionsNetwork Positions

BridgesPeripherals

Group Members

Central MembersIsolate

5

1. At least 10 Other Measures of Centrality & Multiple Algorithms

• Degree

• Closeness

• Betweenness

• Eigenvector

• Power

• Information


2. Different Leaders for Different Stages

100

40

60

80

Per

cen

t A

do

pte

rs

Degree

Betweenness

Closeness

0

20

1 2 3 4 5 6 7 8 9 10

Time

Degree

33. Identify Bridging Nodes. Identify Bridging Nodes

• Leaders are important within groups

• Bridges may be critical to diffusion between groups

• Bridges may be less “burdened” than leaders.

• Bridges may be more amenable to change.

8


14

15

16

19

2

20

21 24

3

30

32

33

37

1

1012

17

18

22

23

25

27

29

31

35

36

4

7

8

11

13 26

28

34

56

99

Two Kinds of BridgesTwo Kinds of Bridges

• Mediators/Brokers (Gould & Fernandez):

C di t–Coordinator

–Consultant

–Gatekeeper

–Representative

– LiaisonLiaison

• Sociometric Bridges: Links, on average, are most critical in the network.

10


44. Identify Isolates or Peripherals. Identify Isolates or Peripherals

• Individuals on the periphery of the network or isolates receive information late or not at all.

• In some settings, isolates may be at increased mental health risk.

11

55. Identify Low Threshold adopters. Identify Low Threshold adopters

• Individuals with low thresholds innovate early.

• They require fewer initial adoptersl

• They are experienced at persuading others.

12


Network Thresholds

= Non= Non--FP UserFP User = FP User= FP User

13

PN Exposure=33%PN Exposure=33% PN Threshold=33%PN Threshold=33% PN Threshold=66%PN Threshold=66%

Low Threshold AdopterLow Threshold Adopter High Threshold AdopterHigh Threshold Adopter

Graph of Time of Adoption by Network Threshold for One Korean Family Planning Community

100% 111

1824293037 4044466616265

Threshold

13

14

15

2

202833

5

52

57

59

60

Time

0%

1963 1973

12 16

22

2531

35

36

414953 64

7

8


Sociometric v. Self Reported OL

• Correlated at approximately 0.43 (p<0.01)

• Self reported OLs less susceptible to peer influence

• Sociometric OLs no more or less susceptible

15

Additional Thoughts about Leaders

1. Many ways to define and identify

2 M t f tl d N t k I t ti2. Most frequently used Network Intervention

3. Many studies of leaders and diffusion

16


Implementation Issues

• Do you just turn leaders loose? How to t t i t tistructure interactions:

– Schedule 1‐1 between leaders & members

– Have leaders give formal presentations

– Have leaders call a meeting

– Allow leaders to decide how to promote changeAllow leaders to decide how to promote change

• Continuum of Passive to Active OL Involvement

17

Strategy Tactic Operationalization

A Taxonomy of Network InterventionsValente, Science, 2012

Identification LeadersBridgesKey PlayersPeripheralsLow Thresholds

Degree, Closeness, etc.Mediators, BridgesPositive, Negative

Proportions, Counts

Segmentation GroupsPositions

Components, CliquesStructural Equivalence, Hierarchies

Induction WOMSnowballMatching

Random ExcitationRDS, OutreachLeaders 1st, Groups 1st

Alteration (Manipulation)

Deleting/Adding NodesDeleting/Adding LinksRewiring

VitalityOn Cohesion, OthersOn Network, On Behavior


2. Segmentation

• Intervention is directed toward or includes a h l f lwhole group of people

• Segmentation interventions identify and expect a whole group to adopt the innovation at the same time

19

2.A. Groups

• Sets of people/nodes that are densely t dconnected

• Groups can reinforce (or inhibit) the behavior change process

• Behavior change may be appropriate for groupsgroups

• There are many network methods used to find groups

20


14

15

16

19

2

20

21 24

3

30

32

33

35

37

1

1012

17

18

22

23

25

27

29

31

35

36

4

7

8

11

13 26

28

34

56

921

2

3

14

15

16

20

30

32

3337

Newman-Girvan 6 Groups

1

4

7

8

12

17

18

1921

22

23

24

25

27

29

31

35

36

22

5

6

9

10

11

1326

28

34


3.C. Match Leaders to Groups

• Rather than have leaders unattached, assign them to people who think they are g p p yleaders

• Leadership is local

• Emphasizes homophily between leaders and members

• Builds on naturally occurring networks• Builds on naturally occurring networks

• Leaders can be more effective if assigned to those who nominate them

23

Network Influence is Proximal

24


4

10 28

32332

4

Sociogram based on ties Optimal leader/learner matching

Matching Leaders to GroupsMatching Leaders to Groups

1

23

5

6

8

10

1314

15

28

29

30

1

3

4

5

68

10

11

12

13

14

28

31

32

33

7

9

11

1231

5

7

9

15

29

30

1(a) 1(b)

Graphical Displays of Intervention Choices


Strategy Tactic Operationalization

A Taxonomy of Network Interventions

Identification LeadersBridgesKey PlayersPeripheralsLow Thresholds

Degree, Closeness, etc.Mediators, BridgesPositive, Negative

Proportions, Counts

Segmentation GroupsPositions

Components, CliquesStructural Equivalence, Hierarchies

Induction WOMSnowballMatching

Random ExcitationRDS, OutreachLeaders 1st, Groups 1st

Alteration (Manipulation)


VitalityOn Cohesion, OthersOn Network, On Behavior 27

Degree Closeness

Keyplayers-Pos.Between


Brokers

Keyplayers

ThresholdsBridges

Groups Positions

Hierarchies


WOM Snow.

Outreach Matching

Vitality Node Addition

Link Deletion Link Addition


Rewire SW Rewire Behavior

Graphical Displays of Intervention Choices


Selecting a NI

• Availability and type of data

– Types of networks

– Existing network structure

• Behavioral characteristics

– Existing prevalence

– Perceived characteristics such as cultural– Perceived characteristics such as cultural compatibility; cost; trialability; etc.

35

• There are several theoretical mechanisms that d i t i d b h i h

Linking Theory to Intervention Strategy

drive contagion and behavior change

• Evidence for a particular mechanisms suggests choice of intervention strategy or tactic


Mechanism Tactic

PowerConflict

LeadersBridges

Influence Mechanisms Aligned with Interv. Choices

ConflictCohesionIsolationThresholds

BridgesKey PlayersPeripheralsLow Thresholds

Group IdentificationStructural Equivalence

GroupsPositions

Information diffusion WOMInformation diffusionHard to reach populationsClosureHomophily

WOMSnowballOutreachMatching

AttributesStructureStructure!!


Leadership & Influence have 2 Dimensions

Trust

Low High

Expertise Low

High


Discussion + Advice Nominations in the South Atlantic (Osteo.)

Prominence Nominations in the South Atlantic (Osteo.)


SNA Application: Analyzing SNA Application: Analyzing Communities by GeographyCommunities by Geography

Average Link Distance


tools for network interventions - inspq · tools for network interventions november 28, 2012...

Documents