The Analysis of Occupational Gender Segregation Over Time and Place: Considerations of Measurement and Some New Evidence

Abstract

Despite the importance of occupational segregation as an area of investigation concerned with understanding women's employment status, pay levels, and promotional prospects during the last 20 years, there has been relatively little attention paid to the problems of trying to measure segregation levels in quantitative data. This paper argues that there are serious measurement problems which it illustrates by showing that two of the principal indices, the widely-used Index of Dissimilarity and the OECD's WE Index are highly flawed and produce unreliable results. It demonstrates the importance of these deficiencies using cross-national data from 9 countries for the period 1970-1982. The paper introduces a new way of analysing the form of these indices in the Basic Segregation Table, which is a 2 x 2 table of gendered occupations by sex. The paper suggests a new approach to measuring occupational segregation which provides more consistent and valid results. This is called Marginal Matching. The paper ends with an analysis of occupational gender segregation in England and Wales from 1951 to 1981. It concludes that, in contrast to research claims to date, the trend in segregation over this period is one of overall stability.