EUSN2019 – 6-hour workshop
This introductory/intermediary workshop aims to offer participants a roadmap in order to appropriately answer their research with social network analysis (SNA). The goal of the 6-hour workshop is to provide an overview and typology of the research questions that can be answered with SNA and to offer a decision tree in order to help identify the type of network data that are needed, and to help decide on the most appropriate statistical technique.
Questions that will guide students include:
(1) Do you need an egocentric or a complete social network approach?
(2) Do I need data from a single network or group, or do I need data from multiple networks or groups? And what is the boundary of the network/group?
(3) Am I looking at the consequences of networks, at the antecedents of networks, or at both?
(4) Do I require a dyadic, individual, group or intergroup approach?
The 6 hour workshop will provide examples of existing research questions, as well as potentially interesting future directions for research at each of the four levels: the dyadic, individual, group and intergroup level.
We first start with a discussion of concepts at these different levels, which will help answer descriptive research questions: the dyadic level (geodesic distance, structural equivalence, …), the individual level (degree, closeness, betweenness centrality, …), the group level (density, centralization, subgroup formation, …) and intergroup level (the centrality of a group among a network of multiple groups).
We then turn to research questions that require statistical methods.
At the group level we focus on how the network structure of a group (its density, centralization and subgroup formation) might impact group outcomes (e.g. group performance), as well as questions about how such structures come about as a result of specific antecedents (e.g. the group’s composition).
At the individual level we discuss research questions, such as why some individuals become more central in a network, and how such network positions might impact outcomes, such as their well-being.
At a dyadic level the focus will be on why a tie between two individuals might emerge. To model the emergence of a tie between two members, the focus will be on statistical techniques such as Exponential Random Graph Models (ERGM) and longitudinal Stochastic Actor Oriented Models (RSiena models and relational event models).
Finally, we will also discuss different ways of dealing with ties across multiple groups and multilevel approaches. The aim of the course is to offer students an overview and guidance on how to answer their research question. Specific techniques on how to use specific software are covered in other courses.
90 Min. Basics and descriptive research questions
10 Min. Break
50 Min. Group level analysis
50 Min. Individual level analysis
10 Min. Break
50 Min. Dyadic level analysis
40 Min. Multilevel approaches
10 Min. Break
50 Min. Intergroup approaches