defining a network

02 Dec, 2025

Migrated from previous Ghost blog in September 2026. This was post 7 of my 2025-26 attempt at the #100DaystoOffload challenge.

I'm learning the basics of social network analysis (SNA). A key reference is the textbook Analyzing Social Networks Using R by Borgatti, Everett, Johnson and Agneessens. I attended their 1-week introductory "summer" school in September too.

Today's tidbit:

..it is the researcher—by choosing a set of nodes and a type of tie—that defines a network.¹

∿∿∿

In network analysis you pick some objects of interest, be they animals or firms or neurons or words, and represent some aspect of their relationship that you're interested in. So these are the fundamental parts of network analysis: the objects (typically called "nodes") and the relations/links (typically "ties").

I fell into a trap that apparently ensnares many SNA novices: thinking that the nodes and ties of a network are "out there" just waiting to be discovered and visualised. Actually, it's a question of research design, according to the 2011 paper "On Network Theory" by Stephen Borgatti and Daniel Halgin, as well as the 2022 textbook from Borgatti and other colleagues.

The position of Borgatti and coauthors is that for SNA you need to define the nodes and ties yourself, very carefully, in close connection to what your research puzzle or topic is, and the way you see the world (your theories about things).

[...] every [date collection] question (such as "Who are you friends with?" or "Who do you seek advice from?") generates its own network, and which to use is determined by the research question.²

There might be a thousand people formally registered as members of Professional Network X. But are they meaningfully connected? It would be uninteresting to visualise that in and of itself. And even then there's an inescapable choice: everyone could be connected to everyone else directly, or via a single central node.

A professional network can have many activities. These could come more or less into focus via different questions, leading to different nodes and ties being chosen:

• If membership of a professional network makes a certain stance or type of activity more likely (perhaps in terms of: how are interactions shaping shared belief/understanding?)
• If shared positions on food policy or shared visions of food system futures predict who collaborates with whom; or the other way around
• If positions within a network (how core vs far away from power) determine level of influence on collaborative outputs
• If external measures of power and prominence (e.g. seniority, revenues/level of funding) predict who is central and powerful in the professional network
• If clusters of collaboration or association are predicted by demographic characteristics, politics, desired food system futures, or shared reference points for evidence about food systems issues

Still trying to work out my exact angle and questions.