[schoen]

This method is explained a little bit at

https://en.wikipedia.org/wiki/NOMINATE_(scaling_method)

I think this may be the first time I've seen this, but I'm quite confused about the claim to be able to make comparisons across time by simple statistical methods. The model has a clear statistical meaning about the similarity of different members of the same Congress, who have all voted on the same set of questions. But surely the issues that the Congress is actually voting on are radically different from decade to decade? Slavery, the Federal income tax, alcohol prohibition, the New Deal?