In 1940, 2,000 women working in a factory were recruited into a study. Half of the women worked in manufacturing and half in administrative offices. The incidence of bone cancer through 1970 among the 1,000 women working in manufacturing was compared with that of 1,000 women working in administrative offices. Thirty of the women in manufacturing developed bone cancer as compared to nine of the women in administrative offices. This study is an example of a: