I am recycling a message that I posted in a Telugu site a few months ago about Simpson's paradox:
For those who do not know (like me recently) I will quicly sketch an example of the paradox. This is supposed to have happenned in the University of Melbourne, but I do not have the actual numbers. The Universitity realized in the 80's that the proportion women academic staff in the total academic staff is less than the Govt. guidelines. Over a period of ten years they increased the proportion of women in every teaching category: tutor, lecturer, reader and professor. After ten years, they found to their horror that even though they increased the proportion of women in every category, in the total academic staff the proprtion of women decreased.
Once one knows this, anybody can work out examples to see that this can happen. So, it seems that one has to be careful with some claims of supposed statistics in newspapers. You can find more examples by google search.

Swarup