Ok, got an outline of an answer from Christopher Jencks on the relation between health & inequality. Wilkinson's original work, he says, was based on data from a small number of countries from relatively few years - but any statistical tests are extremely sensitive to countries, years, control variables, etc. Data for the U.S. shows some correlation, but again, the tests are very sensitive to initial specifications. Since almost every researcher working on the topic comes to it with a strong bias, it's hard to come to a definitive conclusion. More to follow... Doug
