In all countries there are major differences between how genders fare in the jobs market. One of the most evident is that the employment rate for women is lower. According to data from the Active Population Survey (EPA), for example, 67.1% of women in Spain were working in 2017, while the figure for men was 79.2%. In addition, the proportion of women who were working part-time (24.1%) was higher than that of men (7.3%) and, when comparing equivalent jobs, the hourly wage earned by women stood between 20% and 35% below that of men (De la Rica, 2016). 

A percentage of these gender differences in the jobs market can be explained by self-selection among women: under certain circumstances, women may prefer more flexible jobs or those with less responsibility which allow them to combine their work with traditional gender roles. However, in order to understand gender differences in the labour market in all their extension, it is important to ascertain whether, beyond any possible self-selection by women, discrimination also exists in recruitment processes for jobs. A good way of checking this is by conducting research with an experimental focus (so-called field experiments) for the purpose of seeing how employers behave in real situations. 

For this study, fictitious CVs of job applicants aged between 37 and 39 years were sent to 1,372 real job offers in Madrid and Barcelona, and the responses from employers were analysed. For each vacancy, two pairs of equivalent CVs were sent, with the only difference being that, in one case, the applicant was a woman, and in the other, a man. With the rest of the characteristics of the applicants and the job offers remaining constant, any possible differences observed in the responses of employers are interpreted as a sign of gender bias when recruiting. In other words, if with an equivalent CV, women systematically have a lesser probability of being invited to a job interview, it could be concluded that there is discrimination against women. And this is precisely what was observed: out of over 5,600 CVs considered, male applicants were invited to interviews in a higher proportion (10.9%) than women (7.7%). Under equal conditions, the probability of receiving an invitation to a job interview was around 30% lower for women than for men. In the experiment, a heterogeneous group of 18 occupations were considered that varied in their degree of feminisation, and the skills and degree of responsibility they required. Therefore, although the job offers considered were not representative of all the vacancies available in the labour market of Madrid and Barcelona, the results in themselves can be extrapolated to a large group of occupations and workers with the characteristics of the applicants considered. In addition to the characteristics of the applicants, also taken into account in the models were the characteristics of the job offer.

The experiment also enabled confirmation of the possible differential effect on men and women of parenthood and of preparation for the post. For every job offer the researchers sent to pairs of CVs, in other words, four in total. In half of the offers, the applicants were differentiated by gender (two were men and the other two  were women) and by whether they had children or not (one person of each gender had children and the other did not). In the other half of the offers, the applicants also had an equivalent CV except in gender (two men and two women) and in requirements for the post (one person of each gender fulfilled the requirements of the post exactly and the other of each gender was better qualified because they had more work experience, an additional foreign language and prior experience in a post with responsibility). 

As we saw previously, in general terms higher proportions of men were invited to a job interview than women. This happened both for those who did not have children (11.5% of childless men versus 8.8% of childless women) and for those who did (10.3% of fathers and just 6.6% of mothers). However, it is especially noteworthy that gender differences are greater among those with children. The probability of receiving an invitation to interview was approximately 23.5% lower for women without children than for men in identical circumstances, but reached 35.9% lower for women with children with respect to men with the same characteristics. Therefore, the results of the study show that women suffer double penalisation: as women and as mothers. 

With respect to being qualified for the post, we again confirmed that men received more invitations to be interviewed than women, both when comparing CVs of applicants who were suitably qualified for the position (8.5% men, 4.9% women) and when applicants were overqualified (13.1% men, 10.1% women). In this case, the gender penalisation was greater if applicants were less qualified than if, in addition to the required skills, they had knowledge of an additional language and more work experience. This fact suggests that discrimination against women, although it does not disappear, is reduced when applicants are better prepared for the post. The differences observed here on a descriptive level are confirmed with more sophisticated statistical models that consider not only whether applicants were invited to an interview, but also the order in which they received the invitation. With this type of model, it is confirmed that employers give lower priority to women than to men, and that the gender differences are greater when they have children or a lower level of preparation for the post in question.

To understand across all their complexity the effects of gender, number of children and qualification, it is essential to analyse the response rates for different subgroups. The first point that is observed is that gender differences are maintained in all cases, although they are much lower in the subgroup in which women have a higher probability of being invited to an interview (without children and with a high level of qualification); and if there is equality in the CV, the probability of women being invited to an interview is slightly lower (12.4%) than that of men (13.6%). In fact, in this subgroup, the differences between men and women are very small and are not statistically significant. 

To explain these gender biases in recruitment some theories would argue that employers act according to prejudices, whereas others would rather point towards stereotypes. The explanations that point towards prejudices (irrational feelings and emotions about a social group and its members) would suggest that prejudiced employers would be less likely to select women regardless of their skills or work experience (Becker, 1971). However, the data observed do not appear to support this hypothesis. Rather to the contrary, it was observed that the response rate varies according to the personal characteristics of the female and male applicants. This finding suggests the existence of statistical discrimination based on gender stereotypes. In a market with imperfect information, employers apply a rational logic in decision-making, according to stereotypes of productivity or performance of certain individuals or the social groups to which they belong. 

The study reveals that employers alter their decisions when women show that they have greater skills than other similar applicants (mastery of languages, a solid professional career or experience in supervisory tasks) thus contradicting the stereotypes traditionally associated with females such as low productivity, more career interruptions or a lack of supervisory skills. This fact, coupled with the practical absence of discrimination among highly qualified women without children, suggests that the gender biases in recruiting processes are based on stereotypes regarding their productivity, rather than on prejudices (irrational antipathy towards women).

Employers (whether consciously or unconsciously) assess a job applicant by considering the characteristics of the social group to which they belong, whether that of women or of men. In other words, employers hold stereotypes regarding the productivity of women and, in general, tend to see them as less committed to paid employment than men (Bielby & Baron, 1986; Cuddy, Fiske & Glick, 2004). This would explain why gender discrimination in women aged between 37 and 39 years declines when, as applicants in a job recruitment process, they present characteristics different to the stereotypical ones, including greater experience in positions of responsibility and having no children. 

This article has been adapted from these studies:

GONZÁLEZ, M.J, C. CORTINA and J. RODRÍGUEZ-MENÉS (en prensa, 2019): "The role of gender stereotypes in hiring: a field experiment", European Sociological Review.

More bibliographic references:

BECKER, G.S. (1971): The economics of discrimination, 2^nd edition, Chicago: University of Chicago Press.

BIELBY, W.T., and J.N. BARON (1986): "Men and women at work: sex segregation and statistical discrimination", American Journal of Sociology, 91(4).

CUDDY, A.J., S.T. FISKE and P. GLICK (2004): "When professionals become mothers, warmth doesn’t cut the ice", Journal of Social Issues, 60(4).

DE LA RICA, S. (2016): "Las brechas de género en el mercado laboral español y su evolución a lo largo del ciclo de vida", Revista de Ciencias y Humanidades de la Fundación Ramón Areces, 16.