I read Turchin's Secular Cycle where he goes over the data sets he uses to draw these conclusions on earlier societies. While the things he's trying to measure are not easy (for instance he uses the number and dates of found coin hoards as a relative measure of instability in England), the amount of data is pretty impressive. I actually gave up after he did all the data on England and skipped France, the Islamic Empires and I think there was one other society, and went to the conclusion. I haven't read "Age of Discord" yet, but it looks like he has a pretty detailed summary here.

http://peterturchin.com/age-of-discord/