change occupations in order to be distributed to occupational categories in the same percentages as men (or women). The actual levels of sex segregation depend on the level of disaggregation in the categories, with more-detailed categories producing higher values of segregation. Furthermore, studies of detailed job categories show it is quite possible to have nearly complete sex segregation at the level of jobs even though aggregate occupational categories show modest or small levels of segregation of men and women.

Although there were some variations, the levels of occupational sex segregation appear to have been fairly stable, perhaps with small declines, over the first 60 to 70 years of this century. For example, the index of segregation computed on detailed occupational categories for the entire workforce declined from a value of 69.0 in 1910 to 67.6 in 1970 (that is, 67.6 percent of women in 1970 would have had to change their detailed occupation in order for women to have an occupational distribution that matched that of men) (Reskin, 1993). According to Jacobs (1998), by 1990 this index had been reduced to 56.4 percent, and in 1997 it reached 53.9 percent. He further notes that declines since the early 1970s have been the greatest for professionals and managers. The dissimilarity index for college graduates declined 20 points, compared with 11.8 points for those with less than a high school diploma.

To address the question of demographic change within occupations, the committee used data from the Current Population Survey for the 1983 and 1991 periods to analyze the changing demographic and skill composition of the workforce. For example, a particular workforce change over this period—such as the increase in the proportion of the workforce that is black—can be decomposed into across-occupation and within-occupation components.

This decomposition works because the total change in the proportion of blacks can be written as the sum of two components, each of which in turn is a sum. The first component is the sum over all occupations of the change in the proportion of the