The second one was to prepare another major tool for the whole project: a glossary of key terms and concepts. Both tasks were completed. The glossary is composed of 27 entries that cover a broad conceptual terrain. The idea was to create a common platform of understanding across the interdisciplinary research team of concepts which may have different meaning in different traditions or which meaning is not sufficiently clear.
I have not seen the glossary (it is not on the website yet), but I imagine some of our 300 codes relate somehow to one or the other of the 27 glossary terms. So, here’s the idea:
Map ethno codes onto glossary terms. This mapping is of course not based on WP data, but it would be an authorial effort, based on the expert knowledge of the POPREBEL consortium.
This mapping allows us to add glossary terms to the SSN as parent nodes of ethno codes.
Ethno codes are already connected to each other by co-occurrence edges.
Which means we can easily induce a “children nodes co-occurrence network” connecting the glossary terms.
Such a network can be interpreted as the association pattern across the glossary terms as it emerges in the POPREBEL conversation. More collective intelligence, but now the grounded theory approach is combined with the methods deployed to create the glossary.
What connection patterns are the people in the conversation seeing?
Do they resonate with the POPREBEL team?
Are there omissions (important glossary terms that have no children nodes in the conversation – no one is talking about them)? Why?
Are there many important codes that do not map onto any glossary term? Does this imply a divergence between how the conversation vis-a-vis credentialed expert see populism?
Formally, this is the application of a multilayer network approach to semantic social networks. See: 10.4230/DagRep.9.2.1. There is a chapter by me, referred to SSNs as we know them; this would be an extension of SSNs (a new layer in the multilayer network).
If this makes any sense, I am happy to contribute by building a graph in Tulip and throwing it back to you guys for analysis.
Where does this glossary come from? (where are the ‘key terms and concepts’ drawn from?)
I like it as a tool to compare whether these glossary terms have any resonance with the terms generated by the community, if they are for example generated mostly from the academic literature. This will be a very useful thing to do going forward, because it’s a strength of SSNA that it generates terms from the ground-up.
I think it’s a little early for us to do this since the codes are in such a nascent phase, so if we do proceed I would want to wait until we’ve managed as a team to at least code and do one pass of each others’ codebooks to refine our codes. End of March/beginning of April is the earliest I can see this being useful.
I would also be very wary of adding glossary terms as parent nodes without a very clear way of delineating that these codes did not come from inductive work with community member contributions.
Excellent point. For the moment, I do not propose to encode non-ethno data in the OpenEthno database, of course! I can generate the graph in Tulip, as a prototype. If it gives us a major epiphany, we will consider what to do next.