Precis of the Day: Sex Segregation and Statistical Discrimination
Bielby, William T., Baron, James N. Men and Women at Work: Sex Segregation and Statistical Discrimination. The American Journal of Sociology, Vol. 91, No. 4, 759-799.
At the time the article was written, research and organizational level studies indicated that there was a persistent division of labor and occupational job-title segregation along gender lines. Bielby and Baron’s article attempts to reconcile such findings with census occupational data to show that men and women in the same census occupation are sorted into distinct occupational areas or job titles. The authors argue that this is due to demand-side expectational choices by employers and sex differences in skills and the technical requirements of work roles that lead to a sexual division of labor, channeling women across occupations and also to specific job titles and areas within the same occupational roles rather than to “supply-side” factors such as personal choice or socialization.
Empirical research should address segregation among organizations and occupations, and segregation within organizations leading to the differentiation of positions and job titles between men and women. Statistical patterns of segregation are inconsistent with such reasons as “employer rationality” and “optimalization” or employees’ “vocational investments,” and segregation appears to be “built into the hierarchy of organizational positions and is sustained by sex stereotypes and workplace social relations.
The authors relied on data from the California Occupational Analysis Center of the U.S. Employment service between 1964-1979 and the dictionary of Occupational Titles (DOT).
The EEOC classifies occupational groups very broadly in their sex composition surveys, whereas the DOT classification scheme is more particularized. The authors tried to harmonize these disparate classification schemes to better compare segregation within organizations along the lines of job titles and work roles.
The authors assessed occupational measures of segregation across:
Seven major occupational groups
645 detailed occupational categories
Establishment job titles.
The authors used this data to measure the “index of dissimilarity”: the percentage of female or male workers that would have to be reclassified to equalize the distribution of work roles by sex. In addition to this statistical analysis, the authors also relied on qualitative, anecdotal evidence about employment practices in specific organizations.
Bielby and Baron found that formal structures such as unions, specific training requirements, and job characteristics were often implemented by employers according to their gendered expectations of productivity and retention to channel women into more detail-oriented, less physically demanding, less potential for advancement jobs.
Statistical patterns of segregation are not capably explained by supply-side factors as self-selection and unequal human capital. The authors found that occupational segregation within particular organizations is much higher than occupational segregation across organizations. That is, men and women can do the same type of work but be classified by job title and occupational area differently within the same organization, leading to statistical patterns of sex segregation within the same organization. Even when the organization employs both sexes in roughly equal proportions in the same line of work, men and women are usually assigned different official job titles. Small differences in job requirements get amplified into large differences in gender composition. Thus, these statistical patterns of segregation, even in “mixed occupations” and controlling for job skill requirements and human capital, suggest that the reasons for sex-segregation are demand side rather than supply side, and are not fully satisfied by employer justifications of economic efficiency, rationality, or optimality.
This article was interesting, but is of limited utility because it raises questions but does not answer them. The authors admit as much, that the article is intended to question presumptions of market rationality without offering alternate explanations or causes. I would have liked the article to have addressed economic theories of rationality and efficiency more aggressively.