Measuring Discrimination


All modern societies have a high degree of role differentiation and specialization and have conventions for allocating roles and tasks to members on the basis of sex and age, at the minimum, along with a variety of other criteria. The sexual division of labour is sometimes reinforced by formal rules or laws permitting and restricting access to particular occupations, social and political roles. But informal rules and conventions can be equally powerful, and they have persisted long after the introduction of sex and race discrimination legislation in Europe, North America, and other modern industrialized societies in the 1960s and 1970s. Studies of occupational segregation are concerned with the degree of separation between work done mainly by men and work done mainly by women. Theories of occupational segregation claim that women lose out from this separation of male and female workforces. For example, “occupational segregation is a main building block, or the sole foundation-stone of theories of patriarchy. In the USA, race was until recently an important second factor in occupational segregation but it has generally been far less important in the British labour market” (Arrow 1973, pp 3-33).

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!

order now

Policy debates commonly depict occupational segregation as a labour market imperfection or problem, as a reflection of sex discrimination in the labour market and of labour market inequality, as the main source of the sex differential in earnings and of women’s disadvantaged position in the labour market. National equal opportunities agencies have an important role in abolishing overt and covert rules, regulations, and selection procedures that discriminate against women (or men) entering particular occupations. Their official reports necessarily argue against maintaining distinctions between what is regarded as men’s work and what is regarded as women’s work in the interests of creating a more open and competitive labour market. International bodies have identified occupational segregation as the main barrier to women’s full participation in the labour market, and argued for policies that would integrate all occupations.

However, the most vigorous polemics against occupational segregation are found within policy documents from the European Commission. The European Council’s fourth Community action program on equal opportunities for women and men which runs for five years 1996-2000 pinpointed the need for sustained efforts to desegregate the labour market as the principal mechanism for promoting equality between men and women in the labour market. To this end, the EC sought, inter alia, to promote the diversification of vocational choices for women and men; to combat gender stereotyping of skills and jobs; and to encourage women’s entrepreneurship. “It welcomed new research to identify the best and most successful practices, measures, and methods to reduce labour market segregation, to promote the economic independence of women, and to promote the full and equal participation of women in social and economic life” (Bannock & Stanworth, 1990, p 110-16).

The European Commission’s campaign to outlaw occupational segregation is all the more remarkable for refusing to take account of the conclusions of its own research reports, which argue against the desegregation of occupations as the main strategy for improving women’s position in the labour market. Instead, revaluing women’s caring professions and current jobs is argued to be a more effective policy. In effect, comparable worth policies are considered to be a more effective way forward than the unrealistic objective of desegregating the labour market.

Occupational segregation on the basis of sex exists when men and women do different kinds of work, so that one can speak of two separate labour forces, one male and one female, which are not in competition with each other for the same jobs. Segregation can be both horizontal and vertical and no single measure fully captures both these aspects.

Horizontal job segregation exists when men and women are most commonly working in different types of occupation–for example, women are dressmakers and men are tailors, women are cooks and men are carpenters. Vertical job segregation exists when men dominate the higher grade and higher paid occupations and jobs, or when men are promoted further up career ladders within occupations–for example, men are heads of schools while women are teachers. “Occupational classifications vary from 6-10 broad occupational groups to 400-600 specific occupations, and they generally identify some combination of vertical and horizontal segregation” (Bannock & Stanworth, 1990, pp 211-216 ). For example, teachers are usually identified as a separate occupation group, which is staffed by a mixture of men and women, so it is integrated rather than segregated. More detailed classifications distinguish between primary school teachers (dominated by women the U.S. and across Europe), secondary school teachers (with a mixture of men and women), and tertiary level teachers (dominated by men in the U.S. and across Europe). Even more detailed occupational classifications might distinguish between head teachers and directors of educational establishments at each level (who are typically male) and ordinary teachers (who are either typically female or mixed).

The wealth of information contained in occupational classifications has generally been ignored in favor of single number summary measures of the overall level of occupational segregation present in the workforce at any point in time. Statistical indices such as the DI have been seen as facilitating comparisons across time and across countries. Paradoxically, “these studies have regularly produced results that are either implausible, because patriarchal societies such as Japan are found to have the lowest level of occupational segregation while a ‘progressive’ society pursuing egalitarian policies such as Sweden is found to have one of the highest levels of occupational segregation, or else self-defeating, because studies reveal no change at all in occupational segregation in the USA in the period 1900-70 despite dramatic changes in the labour market and occupational structure” (Beatson 1995, pp 152-57). Thus the meaning and validity of these indices must be in doubt. In addition, single number indices are inflexible in their application. The DI, for example, requires comparisons of pairs and only permits comparisons of pairs. It cannot provide a constant measure for comparisons across labour market subgroups, such as full-time and part-time workforces. For these and other reasons, the elegant simplicity of single number indices of occupational segregation has to be rejected in favour of a fuller mapping of the pattern of occupational segregation. The new approach identifies a separate category of integrated or mixed occupations straddling the dividing line between the two dominant categories of male and female occupations.

The characteristics of sex-segregated and integrated occupations are of some importance for theories explaining the persistence of occupational segregation, and for policies seeking to eradicate occupational segregation. It thus predicts that men will always have the major share of the higher-grade and better-paid occupations and that women will always be concentrated in the lowest-grade and worst-paid occupations. Theories developed by economists produce the same predictions, without spelling out explicitly why this outcome is brought about through the mechanisms they describe. Most economic ‘theories’ of occupational segregation do no more than describe the processes through which the pattern of occupational segregation is maintained, such as ‘job queue’ and ‘labour queue’ theories, which say that men (and whites) are always first in the queue and get preferential allocation (or promotion) to the best jobs; ‘statistical discrimination theory’, which states that male employers prefer not to hire (or promote) women because they believe women, on average, to be less committed, productive, stable, and reliable employees; or ‘human capital theory’, which states that men get to the top because they invest in the qualifications and work experience that are rewarded in the labour market, whereas women do not make the same investment. “The only economic theory to explain why these processes occur in all societies is the variety of rational choice theory labeled the ‘new home economics’, which argues that some degree of division of labour is maintained within households because it is efficient and mutually beneficial, yielding productivity increases to the person who specializes in non-market work in the home as well as to the person who specializes in market work” (Dustmann, Micklewright, Rajah ; Smith 1996, pp 79-103). This theory also predicts that men achieve better positions and rewards in the labour market, while most women invest their main efforts and energies in the home and family, but it points out that the roles can also be reversed, equally effectively.

Social science theories have said nothing about integrated occupations because they disappear from view in the dichotomous view of occupational segregation promoted by single number measures of occupational segregation and their focus on ‘male’ and ‘female’ occupations. The few national studies that identified integrated occupations in the past proceeded to ignore them and exclude them from subsequent analyses. Studies that focus on job segregation at workplace level have been even more likely to set aside integrated jobs as too small to merit attention. So there are no theoretical predictions about the characteristics of integrated occupations and limited evidence from rare case studies. Most case studies focus on men and women in non-traditional or sex-atypical occupations rather than on integrated occupations. Since integrated occupations are the ultimate goal of equal opportunities policies, it seems more appropriate to focus the research spotlight on this small group of occupations, to see what positive lessons they might offer to policy-makers, rather than on the negative lessons of sex-segregated occupations. A comparison of the profiles of integrated and sex-segregated occupations is clearly of both theoretical and policy interest (Eardley ; Corden 1996, pp120-31).

Drawing on labour market segmentation theory, it can be hypothesized at the start that integrated occupations would have a special role in labour-market processes of recruitment and job-sorting. The initial hypothesis was that integrated occupations form a kind of no man’s land between well-defined male-dominated and female-dominated occupations, a point of entry to the workforce where people queue to get into conventional and stable segregated and sex-labeled careers. Integrated occupations constitute a busy crossroads, or an antechamber leading into more permanent occupational choices with their associated social and economic profiles. The hypothesis predicted that integrated occupations would consist of gender-neutral jobs with low entry barriers and would be ‘transitional’ jobs for young people entering the labour market at one end of the life cycle, and for older workers phasing themselves gradually out of the labour market into full-time retirement at the other end. Integrated occupations were expected to contrast sharply with the more stable and continuous careers offered by sex-segregated occupations, providing a springboard into these heavily sex-labeled occupations, and being no more than a short-term option allowing new entrants to the labour market coming straight from full-time education, and people (mainly women) re-entering after a long spell out of employment to sort out their options and the opportunities for a longer-term choice of occupation, typically in the segregated sector. Thus the integrated category was expected to consist primarily of relatively low-grade occupations with low entry conditions, high turnover rates, and the most pronounced degree of labour flexibility, however defined, serving as an entry and exit route for occupations in the segregated sector. “This hypothesis was consistent with the small size of the integrated sector, its very slow growth, and would allow some instability in the occupational content of the category” (Eardley & Corden 1996, p 78-82).

It seemed likely that the integrated sector would also include some genuinely integrated occupations offering long-term careers to men and women, but these were expected to be in the minority, even if growing, and greatly outnumbered by the gender-neutral short-term jobs. It is to employers’ advantage to give themselves the widest possible recruitment base for jobs of low skill or with high turnover. This alone suggested that jobs taken on entry to the workforce, and in the retirement years, would be the least sex-labeled of all, thus eliminating any unnecessary restrictions on sources of labour among men and women, young and old. An obvious example would be check-out operators in supermarkets, an occupation requiring limited training and where sex labeling offers no obvious advantage, given that supermarket customers encompass the full range of social characteristics. At the other extreme, personnel managers offer one example of a genuinely integrated occupation with long-term career potential. Most organizations need no more than one personnel manager, who might be of either sex, given that in most establishments the workforce includes men and women. Thus the occupation is open to men and women, even if most personnel manager jobs are segregated by virtue of there being only one or very few such posts in any organization, which might thus be filled by men only or women only.

In the event, integrated occupations did emerge as distinctive in many ways, but not along the lines hypothesized. Highly qualified and genuinely integrated occupations offering long-term careers were found to greatly outnumber low-grade gender-neutral transitional jobs. “In a curious reversal of the initial hypothesis, the large group of female-dominated occupations emerged as the sector containing the largest number of sex-neutral transitional jobs” (Edgeworth, 1922, p 98-104).

It is often argued that women have to be better qualified, and have higher performance on the job, especially in male-dominated occupations, in order to overcome the sex-stereotyping and discrimination that devalues their potential suitability and actual contribution in any job. Women were systematically better qualified than men in all social classes. A cross-national comparative study of employment in the European Union also concluded that, relative to men, women tend to be better qualified for the jobs they hold, especially at professional level. In addition, working women are a self-selected group who are not representative of all women of working age whereas virtually all adult men are employed and it is well established that female work rates increase with the level of educational qualifications held and hence occupational grade. The one exception to this pattern might be female-dominated occupations, but here female over-qualification can again be expected, particularly in part-time jobs, as many women trade earnings and career prospects for convenience factors after childbirth, due to changing their priorities between home and work.

Another distinctive feature of integrated occupations, equally unexpected but consistent with the highly qualified character of the group, is that they have a high proportion of self-employed workers: 20% compared to 17% of people in male occupations and only 3% of workers in female occupations. Female occupations have almost no self-employment; this sector is distinctive in consisting almost entirely of dependent workers. The pattern is consistent across the full-time and part-time, male and female workforces. As a result, the employee workforce is heavily sex segregated, whereas the self-employed workforce is dominated by male and integrated occupations. It appears that self-employment offers an important route to occupational desegregation, “a way of avoiding the well entrenched sex-stereotyping and traditions of the mainstream workforce, where trade unions and employers present social barriers to change and restructuring” (Jacobs, 1989, 160-73).

Half the self-employed in integrated occupations are genuine entrepreneurs or small firm owners who employ others, and half are the solo self-employed who are only marketing their own labour and services and have no employees. This is equally true of men and women. In contrast, almost all the self-employed in male occupations are solo self-employed; very few are small firm owners. This combined with the high percentage of managers and foremen in the integrated sector means that over half of all workers in integrated occupations are in positions of authority, in large or small firms. In this respect, there is a sharp contrast with female occupations, where the vast majority (90%) of workers is dependent employees subordinate to control by others. On this dimension, integrated and female occupations form two extremes, with male occupations in between.

As regards hours worked, the contrast is again between integrated and male occupations, where long hours are typical and where there is almost no part-time work, and female occupations where part-time work (less than 30 hours a week) is common and overtime is rare, with very few people regularly working long ‘overtime’ hours of 46 hours a week or more. The long work hours in male and integrated occupations are partly explained by the presence of the self-employed, who generally work longer hours than employees. But work hours are clearly gender-specific rather than occupation-specific, with a fair degree of consistency in hours among men, and among women, irrespective of the type of occupation they are in: “only 1% of men compared to 21% of women work part-time in male occupations; 7% of men and 43% of women work part-time in female occupations” (Jacobs, 1989, pp 160-73). The sex differential in hours worked is underlined in integrated occupations, where one-quarter of men work very long hours (46 or more a week excluding overtime) while one-fifth of women work part-time hours of less than 30 hours a week, a difference that is likely to have an impact on status attainment and earnings.

Finally the socio-economic status and occupational grade of integrated occupations also establish them firmly as career jobs more often than as low-grade transitional jobs and, beyond that, as a distinctive higher segment in the occupational and class structure.

Integrated occupations are concentrated disproportionately in London and the South-East region of England, which is where the largest concentrations of professional and specialist workers are located. Highly sex-segregated workforces are more common in other regions of England and in Scotland and Wales.

Politicians often argue that growth and prosperity provide a faster route to equality of opportunity than is achieved by affirmative action policies. This is doubtful, at least as regards equal opportunities legislation, given that unequal pay for men and women was maintained over centuries of growth. However, there is some evidence for the USA that economic growth and prosperity provide a faster route to equalizing labour market outcomes than would be achieved by positive discrimination in favour of women or racial minorities.

As regards travel to work patterns, female occupations are distinctively different from male and mixed occupations, but differences derive primarily from the sex differential in job choice which then becomes a characteristic of female-dominated occupations. “Integrated occupations have above average proportions of men and women working at home–6% compared to 2%-3% in the sex-segregated sector” (Bannock & Stanworth 1990, pp 78-85). Despite the popular view that women are trapped into working at home by child care duties, women and men are equally likely to work at home and female occupations do not have the highest rate of home working. The higher level found among integrated occupations results from higher levels of self-employment in this sector, with 19% of the self-employed working at home and another 31% working from home as a base. However, men are more than twice as likely as women to work from home as a base, or to have long journeys to work, while women as a group are more likely to work very locally, with short trips to work, often on foot.

Female occupations are distinctive in being located close to home, due to women’s preference for local jobs offering short journeys to work. Women are almost twice as likely as men to work locally, within a 2-km radius of their home and twice as likely to walk to work. The sex differential persists, to varying degrees, across all three types of occupation. So the travel to work patterns of integrated and segregated occupations is primarily a reflection of their sex composition. Even among part-timers, women are twice as likely as men to get to work on foot. Women’s lower investment, in time and money, in transport to work echoes their lower investment, in time and effort, in obtaining educational qualifications. In effect, employers have structured female occupations to accommodate women’s distinctive preferences for part-time work, local jobs, and occupations requiring little or no investment in training and qualifications. Even on the basis of the limited information available from the 1991 Census, “it is clear that it is female occupations, rather than integrated occupations, that provide the ‘gender neutral’ sector with low entry barriers easing transition into the labour market” (Dustmann, Micklewright, Rajah & Smith 1996, pp 123-19). The female sector is large enough to contain career occupations as well, comparable to the long-term career occupations in the integrated and male-dominated sectors. As a first approximation, about half of the female sector consists of ‘career’ occupations and about half consists of short-term jobs requiring no long-term commitment or investments, and with similarly reduced rewards and opportunities.

Analysis of the personal and household characteristics of workers in each of the three types of occupation again shows that the most important dividing line is between female-dominated occupations and integrated or male occupations. The one-quarter minority of women working in integrated and male occupations differs qualitatively from the three-quarters majority working in female occupations–the former being more career-oriented and the latter being more centered on family and home. Occupational choices reflect the polarization of working women, which is a long-term secular trend rather than a temporary and reversible phenomenon. More surprisingly, the small minority (14%) of men working in female occupations also differ substantively from the majority of male workers.

Women in female occupations are less likely to be single, divorced, cohabiting, or remarried; it is married women who gravitate to female occupations, which have an above average concentration of women in the ‘conventional’ family of two adults with dependent children. Their choice of job reflects greater priority being given to domestic activities than to paid work: they are less likely to work overtime, more likely to work part-time and to travel short distances to places of work close to home. “In contrast, men working in female-dominated occupations are twice as likely to be single and young (aged under 25), still living in their parents’ home, sometimes still in full-time education, and less likely than other men to have dependent children. Like women in these jobs, they also tend to work part-time, in local jobs” (Eardley ; Corden 1996, pp 143-149).

On the basis of personal and household characteristics, integrated and segregated occupations look identical in the aggregate. In all three groups three-quarters of workers live in owner-occupied housing, 4% are non-white, 29% have never been married, and 40% have dependent children living with them in the same household. In the aggregate, the three groups are distinguished by job characteristics, but not by personal characteristics. Another conclusion is that sex-atypical workers are generally young, with older workers reverting to sex-typical occupations.

In sum, the relationship between occupational segregation and earnings only merits further research if datasets can provide information on the full range of social and economic factors that determine earnings, and also employ occupational classifications that are sufficiently detailed to identify the higher earnings of management grades within the increasingly specialized jobs of the Service Class. Such datasets are unfortunately all too rare. The more fruitful focus for future research would be the social processes that lead men to attain jobs of higher status and earnings more often than women. Even in the most enlightened organizations, sex discrimination, sometimes unconscious, is one factor in this process. But that is not the same thing as research results that show sex to be one factor in the process.








Arrow K. J. (1973), “The theory of discrimination”, in O. Ashenfelter and A. Rees (eds.), Discrimination in Labour Markets, Princeton: Princeton University Press, 3-33.

Bannock G., and Stanworth J. (1990). The Making of Entrepreneurs, London: Small Business Research Trust.

Beatson M. (1995). Labour Market Flexibility, Research Series No. 48; London: Employment Department.

Dustmann C., Micklewright J., Rajah N., and Smith S. (1996), “Earning and learning: educational policy and the growth of part-time work by full-time pupils”, Fiscal Studies, 17: 79-103.

Eardley T., and Corden A. (1996). Low Income Self-Employment, Aldershot: Avebury.

Edgeworth F. Y. (1922). “Equal pay to men and women for equal work”, Economic Journal, 32: 431-57.

Jacobs J. A. (1989). “Long-term trends in occupational segregation by sex”, American Journal of Sociology, 95: 160-73.