• based on a conversation with Chris Hoyt

• Assume individuals all have some discrete set of social attributes.
• Social attributes are labels put onto a person by society which influences others' behaviors.
• Each individual can look around in their environment and assesses what social response they would get from any given label
• Most people only care about a few people's opinions, their close friends and family.
  • Some care about welders' respect, or computer enthusiasts'. Any they meet.
  • Some care about an amorpheous body of investors or a ...
• Consider a social group which considers X as "normal"
• As the prevalence of not-X increases in this group, so does the tension they feel.
• This tension can build, compound with other such unusual traits, and generate a real demand for a different "normal".
• This often requires "permission to change" from central and respected figures (stars), but can also come from empathy for a not-X (especially for those who are X).
• Once the most uncomfortable individuals have this permission, they can comfortably reorient their labels.
• This release of tension is seen in every civil rights movement in history on the largest scale.
• The entire story is complicated / made more interesting that what is "normal" is what is common of the often-observed: the media, family, workplace, and other parts of a person's daily interaction with the social.
• Small cork-popping probably happens as well, but I'm not sure what that looks like.
• If this results in the loss in dominance of a group, or in its threat of extinction, it may well radicalize.

Interacting blobs in Blau space

• Consider an N-dimensional space, each dimension represents a social attribute.
• For some variables, the density of people in this space would have some physical properties.
• Groups that were close to each other would actively differentiate themselves from each other. (I could also imagine them joining)
• Clearly distinguishable groups would not need to do this boundary-building, and would freely associate. (I also could imagine them hating each other)
• Important: this description could only apply for certain attributes. Each attribute should have a different "law".
• Identity labels may well obey some simple rules, according to threat, differentiation, and other inherent dynamics of these.

Further notes

Ok, my brain kept running after our call.
I think my own pet topics and bias was still warping the focus of your study.
After hearing you talk about what you were trying to communicate, I have the following recommendation:

• I would (being a marketer) call this "How Social Media Has Accelerated Polarization in Society by Distorting User Perceptions of Collective Behavior."
• While it is a commonly discussed theory that social media exacerbates our desire for selective information (confirmation bias), your unique proposition is that it also deceives us into thinking more people do or believe something to a much larger degree than is true. More uniquely, is that it quantifies it.
• I think you could add, for comparison, a chart that shows how an egonet would reflect reality without social media. It still existed, but at an exponentially smaller effect. It can be assumed, for example, the largest "star" in the egonet would top out at 150 friends (Dunbar's number)
• Yes, TV and Radio hosts still existed, but they don't have as close a connection as "posting relationships" do on social media. The relationships are too distance and lack the engagement to provide the positive reinforcements.
• Of course, the "Satanic Panics" of the late 80s and 90s was due to the large communities of Evangelical churches, and did also create a similar distortion of perpetuating an illusion of events that never actually happened.
• Egonets are more of a term you creating to represent a mathematical fact, it may not be seen as a full hypothesis. - But then again, I don't know how the culture of these things are perceived in the academic world. My viewpoint comes from trying to communicate complex ideas to the general public.