Introduction to Social Network Analysis

Introduction to Social Network Analysis

Format: Course.

Interval: August 6-9, 2019 (4 days).

Level: Introductory, any level (MA & PhD might find it the most useful); no prerequisites needed; previous statistical training handy, but not necessary.

Disciplines: Social Network Analysis.

Course description:

This course provides a hands-on overview of applied social network analysis techniques and their theoretical underpinnings in political and social sciences.

By the end of it, you will be able to independently conduct basic exploratory analyses using different types of relational data and make informed choices about further steps for inferential network analysis and confirmatory analyses in different contexts: politics, economics, sociology, psychology.

We begin with the practical challenges and solutions in working with network data, and then introduce you to network structures, actors’ positions within networks, and the implications of these for different behaviors. Towards the end of the week, we will cover the basics of hypothesis testing and network dynamics.

The course combines workshop-style activities, using participants’ own data and example data sets in any of two preferred software environments (point-and-click or coding), and making use of data visualizations, including discussions of network, political, and social theory in conducting social network research.

The class is not heavy in mathematical formulas, but the basics of network science will be covered and explained through practical examples.

No prior knowledge required.

This course is for those who have heard about network analysis and think it might be a useful toolkit in their own research.

Exploratory network analysis is suitable for anyone doing qualitative, quantitative or mixed-methods research.

Day 1: Working with Network Data

Network data is quite peculiar as compared to typical data for statistical analyses. Their format, storage, and meaning are not always straightforward. Getting the data in the right form for analysis is the most important and often the most time-consuming part of the research. We’ll briefly cover data collection methods, typical database formats, and try some transformation and visualization techniques used in exploratory analyses. We finish with a discussion on diversity of operationalisations and interpretations, using examples from your own work.

Day 2: Understanding Network Structures

The structure of a network can tell you a lot about the underlying relational processes and mechanisms at work. At the macro-level, we explore the different network structures displayed in our diverse empirical data. We discuss what the main network properties tell us about our subject of analysis and do our first network-level analyses: degree distributions, centralization, clustering patterns, communities. You’ll be introduced to the theoretical and technical complexities that span from the results – understanding mechanisms at work in various types of networks. We finish with a discussion on choosing productive avenues for further research based on network statistics at the whole-network level.

Day 3: Understanding Actors’ Positions in Networks

The positions different actors display in the network entail constraints and opportunities for their behavior. We will discuss centrality measures and different theories of relationship formation applied to your research, and explore models for hypothesis testing at the individual level. The central discussion for this day will be the idea of causality in social networks, trying to understand causal pathways to network positions of actors. We finish with a discussion on choosing appropriate avenues for further research based on network statistics at the individual level.

Day 4: Testing Network Hypotheses 

After learning the basics of descriptive statistics in networks, we cover the main techniques for testing network hypotheses at different levels of analysis (macro-level and individual-level): traditional statistical tests, regression models for networks. We end with a discussion on the science and art of choosing the right regression models for networks, assumptions, implications and interpretations of results.

This course covers only basic concepts and analytical techniques. If you come with your data, by the end of the course you will have a first exploratory analysis of your network, as well as a few theoretical leads related to your substantive application. If you don’t come with data, you’ll still be able to conduct a comprehensive exploratory network analysis, and get inspiration for your next research project/thesis/article.

Note   

This is an applied workshop. Please bring your own laptop and have the software installed on your machine. Make sure that they work before coming to class.

Materials about software installation of ORA, Gephi and R, a brief on data formatting, and some R example codes will be available on the course page before the class starts. If you already have a dataset of interest, bring it along. If not, you’ll get access to example networks.

Expect two short practical assignments.

Software Requirements

Please bring your own laptop.

Since the make-up of the group is expected to be interdisciplinary, I will cover two types of software: a point-and-click one (Gephi or ORA-Lite) and a programming language (R).

All software is free, and you are expected to have them installed and working before the course.

Hardware Requirements

Depending on the size of your data and what you want to do with them, generally, the more powerful the computer, the better. 

Course instructor

Silvia Fierăscu

Dr. Silvia Fierăscu Research Fellow Dr. Silvia Fierăscu has a PhD in Comparative Politics and Network Science from Central European University. Her research focuses primarily

Read More »