@Jon and Robert, if you're at all interested, the book is well worth reading. Pinker deals with facts, not correlations--and when he can't get them, he uses multiple correlations to get the best fix on the facts he can. And he has an extensive discussion of how WWII skews his data.
I am probably gonna have a look at his new book. I am familiar with Pinker from some college readings of his earlier works; although he is generally held out as a propagandist, it looks like the critics felt his new book also held some clear bias. I'll hold my judgment till I read it.
My point in my earlier post, as echoed in others, was simply to point out statistics can be misleading. If you understand this constraint leading into a propaganda piece, you can still consume the material knowing the author is navigating you towards a [hopefully] clean and concise theory supported by some evidence. Its still Op-Ed, even if the author is a PhD or Harvard prof, or whatever.
As it happens, I think statistically the US methodology of tracking crime, the defining crime, and quantifying crime are pretty bad. For example, plead-down cases are categorized by the plead, not the act. Or, the exclusion of crimes committed but not reported. I hold similar doubts about other countries.
Ultimately, I get nervous whenever anyone tries to talk me into lowering my guard about anything. For me, I will hear an argument that violence is down when I can leave my house unlocked, my city lowers taxes because we need fewer police, and I don't see a bulletin that 30 people were shot in Chicago over the weekend.