A talk by ecologist Eric Berlow shows how network models and their nodes can help us find simplicity in complex systems, biological or social. The talk and demonstration of the model takes less than three minutes. The presentation is, unfortunately, so brief that he fails to describe the method by which nodes are selected and significant relationships are identified.
What I find most compelling about this concept is the possibility of using network/nodes analysis as a research method. I see that it has the potential to rival or exceed the capacity of statistical analysis to yield useful information for sociologists. The network/node analysis would reveal the significant relationships in a system; statistics would then be used as a refinement to quantify how strong the relationship is.
I completed a course in Structural Equation Modeling earlier this year - I wonder if that's similar to what you might have in mind. I certainly didn't end up as a convert to the methodology, but it does seem to have some good creative potential.
ReplyDeleteAs for this video, I have a bit of a problem with Berlow rather uncritically modeling that US military flowchart, giving the impression that the represented nodes and relationships are an accurate representation of the world. Building these sorts of networks and flow-charts is usually done in an ad-hoc fashion, and my SEM experience at least taught me that the social world rarely conforms to the sorts of causal relationships we try to model. What sort of entities exist in this ontology? What are the sorts of social causal relations at work?
For an even more ridiculous DoD flowchart, the following has to be my favorite:
http://www.wired.com/dangerroom/2010/09/revealed-pentagons-craziest-powerpoint-slide-ever/