gedymin

The first and probably the best, method is to assume that the male and female age-specific death rates from lung, cancer observed among the ACS non-smokers in 1959-72 would have applied to the whole country in 1978 if no one had ever smoked, to work out how many U.S. lung cancer deaths would thereby have been predicted (-12,000), and, since there were actually some 95,000 lung cancer deaths in the United States in 1978, to ascribe the excess (-80,000-85,000 lung cancer deaths) to tobacco.