The division of labour, in the context of paid employment, as a result of which men and women (or members of different ethnic or religious groupings) are channelled into different types of occupational roles and tasks, such that there are two (or more) separate labour forces. It is conventional to distinguish vertical job segregation, by which (say) male or white employees are concentrated in the higher-status and better-paid positions, from horizontal job segregation (where the different sexes or ethnic groups work in different types of occupation—men are engineers, women are typists, and so on). This distinction is discussed further in the entry on the division of labour.
It is important not to confuse occupational segregation (which refers to those processes by which individuals or groups holding particular jobs are kept apart, so that there is little effective competition between them) with labour-market segmentation (the term usually applied to the differentiation of labour-markets into discrete firm-specific segments, each offering different career rewards, conditions of employment, and so forth). There is no necessary relationship between the two, since there are societies where the sexes are separated into different sorts of occupations, but women enjoy relatively good working conditions and career opportunities (at least when compared with other societies in which there is a similar degree of occupational segregation by sex).
Occupational segregation by sex is widespread in all industrialized societies (although it occurs to varying degrees); segregation by ethnic group is more common in the United States than Europe; occupational segregation by religion occurs less frequently (although, for example, to some extent there are still separate labour markets for Protestants and Catholics in Northern Ireland).
Measurement problems loom large in empirical studies of occupational segregation. What proportion of workers in one occupation as against another need to belong to a particular ethnic group in order for us to speak meaningfully of occupational segregation? To what extent must one sex rather than the other dominate the higher-grade and higher-paid occupations before vertical job segregation can be said to exist? One popular measure of horizontal segregation is the dissimilarity index, but various alternative coefficients of both vertical and horizontal segregation are in common use, including for example the coefficient of female representation (which is calculated for each of the major occupational groupings by dividing its female share of employment by the female share of total employment). Many of these indices are sensitive to the degree of detail in the occupational classification that is being used—the more detailed the classification, the greater the degree of occupational segregation that tends to be identified—an important point when comparisons are being made across time and nations (since crossnational standardized occupational classifications are still encountered relatively infrequently in both social surveys and official statistics). Researchers are increasingly aware that any single number index of occupational segregation is unlikely to reveal patterns of change within the occupational structure. One alternative is to measure the differences between integrated and segregated occupations in terms of their relative importance across groups and time. These sorts of measurement problems are analogous to those discussed in the literature on social mobility.