A health study in England surveyed people from 1972 to 1974 and then followed up on them 20 years later. In the original sample, 44% of women were smokers, but only 19% of women over the age of 65 were smokers. At the end of the study, it was found that 24% of the smokers had died while 31% of the nonsmokers had died. Does this mean that smoking is good for you? What would be a better way to look at the effect of smoking on mortality? David R. Appleton, Joyce M. French, Mark P. J. Vanderpump, "Ignoring a Covariate: An Example of Simpson's Paradox", The American Statistician, Vol. 50, No. 4 (Nov. 1996), pp. 340-341.