How to abuse statistics

I heard Harriet Harman on the Today programme this morning discussing pay-rate differentials bewtween men and women and she came up with this glorious piece: "Female part-time workers still earned 40% less per hour than their full-time male counterparts" (http://news.bbc.co.uk/1/hi/uk_politics/7474801.stm) That is a load of sphericals to be honest. Try comparing like with like Harriet: part-time male workers with part-time female workers, or perhaps female full-time workers with male full-time workers?