Network Analysis

Prof. Henry was always willing to go the extra mile and explain everything very clearly. His course provides an excellent introduction to network analysis. I learned a lot! — participant from Hong Kong

Network concepts are increasingly prevalent across a wide range of disciplines and are often used as a tool to study complex phenomenon, such as cooperation, diffusion of innovation, and social capital. This course introduces participants to major research questions in the study of networks and is concerned with basic methods commonly used to analyze networks in the social sciences. It covers such topics as the managing of network data, visualization of networks, and descriptive network analysis. If class circumstances permit, higher levels of course content may be included, such as models of network evolution and advanced methods for drawing statistical inferences about network structures and the behaviors of network actors.

Although the study of networks is theoretically motivated, this course is primarily focused on developing practical skills in working with network data. Participants learn, through hands-on training in the free software package R, how to manage and rigorously analyze network data.

Dates

This one-week, 35-hour course runs Monday-Friday, July 1-12, 2024. The course is scheduled for 9.00am-12.30pm.

Instructor

Adam D. Henry, University of Arizona

Detailed Description

This course begins with a theoretical examination of networks and their importance in the social sciences and discussion of essential research questions that motivate the study of networks, including:

How do various types of networks influence social, political, and economic problems of interest?
How do networks self-organize and evolve over time?
What contextual or institutional factors influence how networks evolve and influence problems of interest?

Although the course is theoretically motivated, it primarily focuses on developing the basic skills of working with network data, network visualization, and running descriptive analyses of network data. It covers descriptive network analysis techniques that allow participants to measure the positions of individual actors (or 'nodes) within a system, identify cohesive subgroups, as well as analyze characteristics of entire network structures. Participants learn about concepts of network centrality, clustering, community structure, and network segregation, and how these concepts can be used to capture variables of theoretical interest in the social science, such as social capital and brokerage.

Through hand-on training in R and the use of real-world datasets, participants acquire practical skills that allow them to create and interpret network graphics, empirically measure network structures, and answer such questions as who is relatively central, who is relatively peripheral, what distinct communities exist within a larger structure, or is a system relatively fragmented or integrated?

After a brief review of topics related to managing network data, network visualization, and descriptive network analysis, this course explores models of network growth and evolution, applications of these models to real-world networks, and techniques used to make inferences about network structures. Inferential network analysis differs from traditional statistical inference techniques because they allow the analyst to explicitly account for the interdependence between nodes and links in a network.

The course covers the following topics:

Random graph models and the application of random graphs to understand how real-world networks grow and evolve over time,
How to quantify the similarity and differences between networks using the quadratic assignment procedure (QAP) and multiple regression quadratic assignment procedure (MRQAP), and
How to estimate the processes that generate networks, such as homophily (attraction to those who are similar) and transitivity (where friends of friends become friends) using exponential random graph models (ERGM).

Moreover, this course provides training in related topics and skills, including programming in R as well as the use of agent-based models to understand network evolution.

Prerequisites

There are no formal prerequisites. It would be beneficial if participants had some experience with R. However, even participants unfamiliar with this statistical software will be able to effectively participate in the course.

Requirements

Participants are expected to have access to an internet-connected computer. Access to data, temporary licenses for the course software, and installation support will be provided by the Methods School.

Core Readings

Scott, John. 2000. Social Network Analysis: A Handbook. 2^nd edition. Thousand Oaks, CA: Sage Publications.