| Yes and: Ezra Klein provides a lot of context in the book Why We Are Polarized. It was long argued that partisanship (Dem v GOP) should become polarized, so that voters knew what they were getting. (Careful what you wish for.) Sorting proceeded polarization, then became self reinforcing. To your point about identity, somehow sorting and polarization led to a stacking of our identities, creating super identities. So that any given belief becomes strongly associated with many others. Like a NASCAR fan is more likely a GOP voter. This is somehow like homophily and schismogenesis. Klein concludes by punting on how to unwind polarization. Instead, he advocates majoritarian rule by ending the vetocracy. So then our polarized parties are better held accountable. -- This paper's Attraction-Repulsion Model is worth considering. I really like their use of dynamic simulations. Coolness. Alas, I currently have no clue if ARM has any predictive power (real world use cases). To noob me, sorting and polarization -- tribes and wedge issues -- kinda look similar to k-means clustering and bee colony optimization. ARM isn't so different. I've been reading about voter behaviour modeling. Starting with William McPhee's work. Again, I have no idea if what they're actually doing is any better than phrenology. Having worked on campaigns, I actually know the mechanics of polling, profiling, GOTV. But the work of orgs like Data for Progress, fivethirtyeight, and our local party's quants has always struck me as black magic. (Or complete bullshit. I still haven't decided.) |