SNA Theory and Design

Introduction

Social Network Analysis: a theory and set of methods focused on the meaning of connections and social structure.

  • The point of SNA
    • Relationships, and how we connect with one another, matter!
    • More so than individual traits or characteristics*
    • The way networks are patterned and structured also matters

Important Terms

  • Network - A group of individual entities connected in a meaningful way
  • Node/Actor/Agent - Individual units
    • People
    • Organizations
    • States
    • Proteins
    • Nuerons
  • Edge/Tie/Connection - Defined relationship or connection between nodes
    • Directed or undirected
    • Reciprocal or not

Attributes vs. Relations

  • Attributes: What we measure all the time!

    • Income
    • Education
    • Gender
    • Self-efficacy
    • Behavioral variables (e.g., physical activity)
  • Relations: ties and structures within networks

    • Who do you know, talk to, trust, spend time with, etc.
    • How closely are you connected to others? How many people are you connected to?
    • Is the network you’re apart of dense, hierarchical, clustered and does that matter?

Basic Assumptions that make SNA different

  • Independence is NOT assumed
  • Actually, that’s an irresponsible way to think, according to network theory
  • “The whole is more than the sum of its parts”
  • Nonlinearity
  • Inputs and outputs
  • Variance explained

Why Might We Need SNA?

  • Dissatisfaction with attribute theories of behavior
  • “Qualintative”
  • More realistic modeling of human behavior
  • Behaviors and diseases spread through social contacts, so model that
  • Develop better programs/interventions

Why Might We Need SNA?

  • It’s SUPER interesting!!
  • The field is growing and continues to be “written”
  • Applies across physical, biological, and social sciences

Two Approaches to SNA

Egocentric Network Research

  • Focuses on personal networks of individual people
  • The ego is the “hub” of the network
  • Constrained by the environments and activities in which the ego is embedded
  • Fits well within standard social/behavioral research

Egocentric Network Measures

  • Composition
  • Homophily
  • Heterogeneity
  • Structural Holes

Whole Network Research

  • All sets of ties among all members of a given network are studied
  • All alters in a whole network are egos, and all egos are alters
  • No longer a focal ego
  • Allows for individual, group, and network level analysis

Whole Network Measures - Centrality

  • A property of a person’s position in a network
  • Where does someone “land” in relation to other nodes in a network?
  • Central nodes usually carry positions of popularity, power, and prestige
  • Centrality typically implies structural importance
  • Central nodes often have influence in behavior spread across a network
  • Several measures of centrality
    • Degree
    • Betweenness
    • Eigenvector Centrality
    • Closeness

Whole Network Measures - Group-Level

  • Subset of a network
  • Component (most basic): all nodes that can reach one another through any number of steps; nodes that cannot reach one another are in a separate component of the network
  • K-core: subset of the network in which each node is connected to at least K other people
  • Creates a density factor for groups
  • Clique: all members of a group are connected to all members of that group
  • SNA posits that people who engage in a particular behavior are often surrounded by other people who also engage in that behavior, or at least approve of doing so

Whole Network Measures - Network-Level

  • Calculated on the whole network (as opposed to each node)
  • Investigates the network from a global (or bird’s eye) perspective
  • Density
  • Centralization
  • Average path length
  • Density and Centralization
Previous
Next