UNIT III
Network topology and diffusion, Contagion in Networks,
Complex contagion, Percolation and information, Navigation
in Networks Revisited.
1. Network Topology in Social Network Analysis (SNA)
Network topology refers to the structure or pattern of connections
among nodes (individuals, groups, or organizations) in a social network.
It describes who is connected to whom and how they are connected.
Key Topologies in SNA:
Centralized Network:
o Few nodes act as hubs with many connections.
o Information flows quickly but is vulnerable if central nodes
are removed.
o Example: A manager in an organization.
Decentralized Network:
o Multiple hubs exist.
o Information is distributed more evenly.
o Example: Communities with several leaders.
Distributed Network:
o All nodes are relatively equal in connections.
o No single point of failure.
o Example: Peer-to-peer online communities.
Small-world Network:
o Most nodes are not directly connected but can be reached
through a small number of intermediaries (“six degrees of
separation”).
Scale-free Network:
o Few nodes (influencers) have many links, while most have
few.
o Common in real-life social media platforms (Twitter,
Instagram
Measures in Network Topology:
Degree Centrality – Number of direct connections.
Betweenness Centrality – Control over information flow.
Closeness Centrality – How quickly a node can reach others.
Density – Proportion of actual ties to possible ties.
Clustering Coefficient – How interconnected a node’s neighbors
are.
Diffusion in Social Network Analysis
Diffusion refers to how information, behaviors, or innovations spread
through a network.
Types of Diffusion:
1. Information Diffusion – Spread of news, rumors, or knowledge.
2. Innovation Diffusion – Adoption of new ideas, technologies, or
products.
3. Behavioral Diffusion – Spread of habits, norms, or practices.
Models of Diffusion:
Contagion Model:
o Spread is similar to infectious diseases (e.g., viral videos).
Threshold Model:
o Individuals adopt behavior once enough peers have done so.
Cascade Model:
o A single action triggers a chain reaction (e.g., retweet
chains).
Factors Affecting Diffusion:
Network Structure – Dense vs. sparse networks.
Influencers (Opinion Leaders) – Highly connected or respected
nodes.
Tie Strength – Strong ties spread trust; weak ties spread new
information.
Homophily – Similar people are more likely to influence each
other.
Contagion in Networks
Contagion in Social Network Analysis (SNA) refers to the way ideas, behaviors,
information, innovations, or even diseases spread across a network of connected
individuals or organizations. It borrows from epidemiology (study of disease
spread), but is widely applied to social, economic, and organizational contexts.
1. Meaning of Contagion in Networks
Contagion = the process by which something “catches on” and diffuses
from one node (person/organization) to another through social ties.
It explains how influence flows and why certain behaviors or information
become widespread.
2. Types of Contagion
1. Simple Contagion
o Transmission requires only one contact.
o Example: catching a virus, hearing a rumor once, receiving a piece of
news on WhatsApp.
o Model: Similar to infectious disease models (SIR, SIS).
2. Complex Contagion
o Adoption requires multiple sources of reinforcement before an
individual changes behavior.
o Example: adopting a risky behavior, joining a protest, buying new
technology (needs peer validation).
o Model: Threshold models – people adopt only if a certain fraction of
their neighbors already adopted.
3. Mechanisms of Contagion
Social Influence: People conform due to peer pressure, norms, or desire to
fit in.
Social Learning: Individuals observe outcomes of others’ actions before
deciding.
Homophily + Network Structure: People with similar interests cluster,
reinforcing contagion.
4. Factors Affecting Contagion
Network Topology:
o Dense clusters slow simple contagions (redundant ties).
o Weak ties help spread new ideas across communities (Granovetter’s
"strength of weak ties").
o Bridges and hubs accelerate diffusion.
Threshold of Adoption: Number/percentage of neighbors needed for
influence.
Tie Strength: Strong ties = trust and reinforcement; weak ties = access to
new information.
5. Models of Contagion in Networks
Epidemic Models: SIR (Susceptible-Infected-Recovered), SIS (Susceptible-
Infected-Susceptible).
Threshold Models: Granovetter’s Threshold Model, Watts’ Global Cascade
Model.
Independent Cascade Model: Each infected node has a probability to
influence its neighbors.
6. Applications
Epidemiology: Modeling disease outbreaks (COVID-19 spread in
communities).
Marketing: Viral marketing, product adoption.
Politics: Spread of political movements, protests, ideologies.
Technology: Adoption of innovations like smartphones, apps, or social
media platforms.
Finance: Contagion in economic crises (defaults spreading across banks).
Complex Contagion in Social Network Analysis
1. Basic Idea
In social networks, contagion means the spread of something (like ideas, behaviors,
innovations, or diseases) from person to person through connections.
There are two main types:
o Simple contagion → Spreads like a virus: even one contact can be enough (e.g.,
flu, rumors).
o Complex contagion → Needs reinforcement from multiple contacts before
adoption happens (e.g., adopting risky behavior, joining a protest, changing
lifestyle).
2. Definition of Complex Contagion
Complex contagion occurs when an individual requires exposure to the same behavior, belief, or
innovation from several different sources before adopting it.
Adoption threshold is higher.
Example: A person may need to see several friends going to the gym before they
themselves start.
3. Key Characteristics
Threshold effect: People adopt only if the number (or proportion) of neighbors who have
adopted exceeds a threshold.
Social reinforcement: Adoption depends on repeated signals from different sources, not
just one.
Collective behaviors: Often linked with movements, social norms, or cultural shifts.
4. Examples in Real Life
Joining a political protest (people need assurance from multiple peers for safety and
credibility).
Technology adoption (new apps become popular only when many peers are already
using them).
Health behaviors (like quitting smoking, using seatbelts, vaccination decisions).
5. Importance in Social Network Analysis
Helps explain why some ideas spread slowly but widely, while others fail.
Useful in modeling diffusion of innovations and behavioral changes.
Identifies network structures that support or block contagion:
o Dense clusters → support complex contagion (strong reinforcement).
o Sparse ties → better for simple contagion (fast spread).
6. Mathematical / Model Perspective
Often modeled using threshold models (Granovetter’s threshold model, Watts’ model).
Each node adopts behavior if fraction of neighbors ≥ threshold.
