Reading through this post, I was struck by how differently you illustrate Hallin’s spheres from him. In particular, while Hallin assumes that the sphere of legitimate controversy is the smallest one, while your normal distribution assumes that most people and conversations are within the sphere of legitimate controversy.

I wish public reactions to people who do predictive modeling hadn’t done so much to lower the public perception of the value of representative opinion sampling. Because I suspect we’re going to need to work harder to ground our understanding in better evidence of the public’s actual views, beliefs, and attitudes in the coming years.