It isn't enough to show that the data is wrong, but it is enough to keep you from drawing solid conclusions from the data.
I disagree, Tim. I suspect there is a literature on this very issue that could be accessed, that would tell us just how important these factors are. And since I begin with the assumption that Krugman is smart and has available to him at Princeton what used to be several very good demographers who could answer these questions with one phone call and/or also has available New York Times fact checkers, I'm assuming he's aware of that literature, has taken it into account but doesn't have room for a great detail in a short op ed column.
So, I would wish to see, not only the argument, with data, that Krugman should have included your factors, but, if possible, the larger literature that discusses this issue. All of that before I would be willing to conclude he was wrong on this point.
Again, I don't expect you to do any of this. Just putting my criteria on the table.
You may recall the discussion Bill and I had about openly carrying weapons. Bill argued, as did Mike (uw), that so doing lowered crime rates. Bill called up Lott's research to substantiate that. I thought that was a terrific way to have the conversation. That led me to dig up the research literature, one rather critical piece of which argued (a) using Lott's data and methodology, the crime rates went up, or (b) the data was simply not strong enough to support either thesis.
That's good public policy debate for my money. |
