Please wait...
Advanced Methods for Social Network Analysis
About
Location:
Mannheim, B6 4-5
Mannheim, B6 4-5
Course duration:
09:00-16:00 (CEST / UTC+2)
General Topics:
Course Level:
Format:
Software used:
Duration:
Language:
Fees:
Students: 550 €
Academics: 825 €
Commercial: 1650 €
Additional links
Lecturer(s): Lorien Jasny, Laura Roldan Gomez (Teaching Assistant)
Course description
Social network analysis focuses on relationships between or among social entities. In studying social networks, we often use many different descriptive analyses, like which actors are the most central or how clustered the network is. This course moves beyond these descriptive measures and presents an introduction to advanced concepts, methods, and applications of social network analysis used in the social and behavioral sciences focusing on statistical methods for social networks. Often, traditional statistical methods cannot be used for the analysis of networks because their relational nature violates the independence assumptions of standard approaches. In this course, we will discuss where and when standard approaches can be used and why they often can't. We'll then cover some of the standard statistical models for networks including Network Regression, the Quadratic Assignment Procedure, the Network Autocorrelation Model, Exponential Random Graph Models (ERGM), and the Stochastic Actor Oriented Model (SAOM). We will also focus on approaches for multilevel, multimode, and temporal data.
For additional details on the course and a day-to-day schedule, please download the full-length syllabus.
Organizational Structure of the Course
Each day will consist of lectures in the morning which often will include demonstrations of code to run the models discussed in the lecture. The afternoon will then include lab or group discussion time. During this period, participants will be given exercises to work through on their own or in small groups to apply the methods described in the morning. Both lecturers will be available to assist as needed. Exemplar datasets will be provided, but participants are encouraged to bring data they are interested in exploring to the course. Participants can also work on their own projects during this time as well as present their projects to the lecturers and/or other participants for feedback and discussion. We will often conclude with a short group discussion of the main themes of the day.
Target group
You will find the course useful if:
- You have already had an introduction to social network analysis and want to move beyond descriptive measures to start to develop and test hypotheses about network data. Or, you have the idea that one (or more) of the models we'll focus on might be useful in your research but you'd need some guidance in getting started and understanding the practicalities of applying these models. Or, you understand the basics of the models you want to use in your research, but your data are more complex (missing data, limitations from the data collection process, multilevel or multiplex ties, or longitudinal) than standard approaches accommodate.
Learning objectives
By the end of the course, you will:
- Understand why the relational nature of network data violates assumptions of standard statistical models
- Be comfortable with the major R packages for social network analysis (Statnet and igraph)
- Be able to generate different hypotheses for social network data
- Know how to apply a variety of statistical models for networks
- Understand how the Exponential Random Graph (ERGM) and Stochastic Actor Oriented (SAOM) models work and what the differences between them are
- Know how to apply these models to advanced data structures like multilevel, multimodal, and temporal network data
Prerequisites
- Be comfortable with basic R programming and data management, i.e. know how to load your data into R, understand how to work with data frames and matrices to generate basic descriptions or plots of the data contained within, and understand the differences between bracket and dollar sign notation.
- Familiarity with descriptive measures from social network analysis like centrality, centralization reciprocity, transitivity, homophily, structural equivalence
- Understand basic statistical metrics like correlation, t-tests, linear regression, logistic regression and how to perform these tests in R
Software and Hardware Requirements
Participants should bring their own laptops for use in the course.
The R statistical programming package is required to follow along with the course material and is available as a free download (http://cran.r-project.org/). We also suggest downloading the free software program R Studio (http://www.rstudio.com), which offers a user-friendly interface to R. In addition to the base packages that are installed with R, we will also use the following additional packages: statnet, igraph, ndtv, numderiv, coda, nlme, and trust. If possible, please install them prior to the start of the course.
