Sage Journals: Discover world-class research

Abstract

Relying on data from a large-scale field experiment in Sweden, this article studies discrimination in recruitment on the basis of gender and ethnicity combined with recruiter gender. The study includes 5641 job applications sent in response to advertised vacancies, and the employer callbacks to these. Gender and either a Swedish or a foreign-sounding name were randomly assigned to the applications, and recruiter gender was documented whenever available. Based on the callback rates, there is evidence of ethnic discrimination against foreign-named job applicants by both male and female recruiters. Also, male applicants with foreign-sounding names are discriminated more than female applicants with foreign-sounding names. Thus, the results show gendered ethnic discrimination in the Swedish labor market, and this does not appear to depend on recruiter gender in general. However, the patterns for gendered ethnic discrimination by recruiter gender vary across occupational categories.

Keywords

Discrimination ethnicity field experiment gender labor market recruitment

Introduction

Labor market inequality takes many forms, and different groups of individuals experience a range of challenges in the labor market. Natives generally have an advantage over immigrants in the labor market, and there is evidence of ethnic discrimination in recruitment both internationally and from Sweden (cf. Baert, 2018). Also, men tend to have an advantage over women in the labor market (e.g. Charles, 2011). In general, men have higher wages and hold authority positions to a greater degree compared to women (Grönlund et al., 2017). These disparities may be partly because of differences in the work- and family-related decisions made by men and women. Yet, it has been suggested that discriminatory behavior by employers may be one possible reason for the observed gender disparities in career-related outcomes (e.g. Blau and Kahn, 2017; Mandel and Semyonov, 2006).

While there is some evidence of gender discrimination against women in recruitment internationally (e.g. Goldin and Rouse, 2000; González et al., 2019; Riach and Rich, 2006; Weichselbaumer, 2004), Swedish field experiments do not find such evidence in general (Bygren et al., 2017; Bygren and Gähler, 2021; Carlsson, 2011). Also, some recent evidence from Europe shows female applicants to be slightly advantaged over male applicants in recruitment (Birkelund et al., 2019; Di Stasio and Larsen, 2020). When it comes to the intersection of gender and ethnicity as regards discrimination in recruitment, there is evidence of ethnic minority male applicants facing more discrimination than ethnic minority female applicants in Europe (Arai et al., 2016; Dahl and Krog, 2018; Liebkind et al., 2016; Midtbøen, 2016). Yet, not all studies find such a difference, and it may differ depending on the type of occupation (Bursell, 2014). Moreover, a previous Swedish field experiment shows that the gender of the recruiter together with the occupational type is important for male and female applicants’ chances of getting a positive response when applying for a job (Erlandsson, 2019).

While a lot of the field experiment literature on gender, and ethnic, discrimination has a theoretical focus on statistical and taste-based discrimination, and occasionally on status characteristics, additional theoretical approaches, such as homophily, in-group bias, and intersectional theories, may also be of importance. Yet, data limitations and endogeneity often make it difficult to study in-group bias based on gender in the real world outside the laboratory (Sandberg, 2018). Studying in-group bias related to several discrimination dimensions, that is, gender interacted with ethnicity, can therefore be even more difficult in practice and is related to increased theoretical complexity.

To my knowledge, there are no previous field experiments on ethnic or gendered ethnic discrimination that focus on the gender of the recruiter, except for one that is limited in scope (Edo et al., 2019). Thus, it is unclear what role, if any, the gender of the recruiter plays in (gendered) ethnic discrimination in the first stage of the recruitment process, that is, in screening job applicants for vacancies. The void in this area of research leads to the following research questions: Do male and female recruiters discriminate against job applicants with foreign-sounding names? Does this tendency differ based on the gender of the applicant? Are there differences in ethnic majority and ethnic minority job applicants’ chances of getting a callback, that is, a positive response to their job applications, based on the gender of the recruiter, and based on the gender of the applicant? Ethnicity here is signaled by typical Swedish names and foreign-sounding names – common Arabic or Slavic names – which is a common practice in field experiments on recruitment (cf. Gaddis, 2018), and may be taken as an indication of immigrant background.¹

Labor market discrimination based on ethnicity and gender, and any combination of these, together with the recruiter gender, is an important topic of study. This is not only because of the scarcity of research on the topic, but also because of the intersection of several group identities or group characteristics, here gender and ethnicity, may entail different, more or less strong, grounds and forms of discrimination (cf. Ridgeway and Kricheli-Katz, 2013), not only as such but also when considering recruiter gender. Moreover, if gendered ethnic discrimination differs by recruiter gender, it may have implications for gender and ethnic segregation in the labor market, given that the proportion of female and male recruiters differs between occupational categories. If recruiter gender plays a role in (gendered) ethnic discrimination, it suggests that gender diversity among recruiters within a workplace is important for recruitment outcomes.

Sweden, together with other Nordic countries, tends to be ranked high in terms of gender equality in labor market outcomes, for example, employment rates (OECD, 2017; United Nations Development Programme, 2016). While the Swedish institutional setting supports a dual-earner model, and the country has a high prevalence of gender egalitarian attitudes (Edlund and Öun, 2016), the Swedish labor market is characterized by a high degree of gender segregation (European Institute for Gender Equality, 2017). At the same time, Sweden has received a rather high number of immigrants, at least in comparison to its Nordic neighbors. Consequently, Sweden has a diverse population with every fifth person being born abroad (Statistics Sweden, 2020). Many of the immigrants have arrived in Sweden because of humanitarian reasons, often without basic education or facing trouble in transferring their qualifications, and the employment gaps between natives and immigrants are among the largest in the Organisation for Economic Co-operation and Development (OECD), across the educational distribution (OECD, 2014: 5). Thus, Sweden provides an interesting context to study recruitment discrimination based on the combination of gender and ethnicity together with recruiter gender.

Literature review

The term homophily is common in sociology when referring to the tendency of individuals to select relationships with similar others (e.g. Kandel, 1978; McPherson et al., 2001). In social psychology, the concept of in-group favoritism (e.g. Tajfel et al., 1971) or in-group bias (Hewstone et al., 2002) is applied to discrimination; research finds that individuals discriminate in favor of those who belong to their in-group over individuals of other groups, and gender represents one basis for such group division (cf. Brewer and Kramer, 1985; Hewstone et al., 2002). In line with the ideas of homophily and in-group favoritism, male employers are expected to prefer male job applicants over female applicants while female employers are expected to favor female applicants over males. Also, one may expect Swedish employers, in general, assuming that most of them represent the ethnic majority,² to favor ethnic majority job applicants over ethnic minority applicants, that is, job applicants with a foreign-sounding name.

The economic theory of labor market discrimination is generally divided into statistical discrimination (Arrow, 1973; Phelps, 1972) and taste-based discrimination depending on the underlying source of discrimination (Becker, 1971). A main difference between these two theories is that taste-based discrimination arises from an individual's prejudice against, or a negative “taste” for, members of a certain group, for example, characterized by ethnicity or gender (or both), whereas statistical discrimination is based on rationality about the average productivity of members of a certain group, such as men and women, or foreign-born and natives when the actual productivity of the individual is unknown.³

For instance, statistical discrimination against job applicants with foreign-sounding names can arise because of average differences in (Swedish) language skills, implying that immigrants, and workers with foreign names, are less productive and thereby more costly to the employer than native workers, for example, workers with typical Swedish names. Also, female workers (of childbearing age) in general, rather than male workers, may experience statistical discrimination because of a lower productivity associated with motherhood. This is possible because female workers are likely to either have children or be perceived as being “at risk” of having children, and employers may link (potential) motherhood to costs such as absence from work and reduced working hours.

Yet, discrimination can arise from a cognitive bias functioning beyond the productivity of certain groups (cf. Correll and Benard, 2006). Status characteristics related to gender and ethnicity can be relevant in such a situation. Culturally prevailing status rankings related to the characteristics of certain groups based on for example, gender, are often used to evaluate individual behavior and characteristics, and according to such status beliefs, men tend to be seen as superior to, and more competent than, women in most domains of social life (Ridgeway, 2011; Ridgeway and Smith-Lovin, 1999). A main distinction between homophily, or in-group bias, and the status characteristics theory is that status characteristics are beliefs possessed by everybody, in general, including the out-group members, meaning that both female and male recruiters would rely on the same status rankings and, according to these, prefer individuals of the superior group, for example, men over women, in recruitment. Whereas if homophily, or in-group bias, based on gender prevailed, only male recruiters, but not female recruiters, would favor male job applicants over female applicants. However, the context is important for the relevance and influence of status characteristics. One may argue that in relatively gender-neutral contexts (e.g. gender-balanced occupations) status beliefs would imply that men, in general, have a slight advantage over women, in masculine-typed situations (e.g. male-dominated occupations) men would be expected to have an advantage over women whereas, in feminine-typed situations (e.g. female-dominated occupations), men's advantage over women would be smaller or even reversed (Ridgeway and Correll, 2004). Also, given that foreign-born are often concentrated in low-paid occupations, employers may even prefer some groups of immigrants for certain low-paid jobs due to, for example, a perceived higher work commitment and job-specific skills (Friberg and Midtbøen, 2018), one may assume more ethnic discrimination in highly qualified occupations in comparison to less qualified occupations.

Yet, the presence of multiple simultaneous group identities or group characteristics, for example, gender interacted with ethnicity, tends to lead to theoretical complexity. Cultural beliefs and stereotypes related to gender and ethnicity illustrate the views of dominant groups in society and influence the expectations about individuals representing different groups (Ridgeway and Kricheli-Katz, 2013). However, intersectional theories propose somewhat competing approaches.

First, the multiple burden (cf. King, 1988) or double jeopardy approach argues that ethnic minority women experience more discrimination, in relation to ethnic minority men as well as ethnic majority men and women, by belonging to two subordinate groups simultaneously (Berdahl and Moore, 2006; Nelson and Probst, 2004; for a review, see Purdie-Vaughns and Eibach, 2008).

Second, and in opposition to the first perspective, Purdie-Vaughns and Eibach (2008: 377) argue that having several subordinate-group identities (an ethnic minority female) makes a person “invisible,” or off-diagonal, in comparison to those with only one subordinate-group identity (an ethnic minority male) as all of a person's group identities will not be fully recognized. This suggests that if an ethnic minority man is the prototype of an ethnic minority, then ethnic minority women can be affected by intersectional invisibility and face less discrimination than ethnic minority men.

Third, the subordinate male target hypothesis (SMTH), derived from the social dominance theory, states that ethnic minority men experience more discrimination than ethnic minority women because of a greater perceived threat (Sidanius and Pratto, 2001; Sidanius and Veniegas, 2000). As the SMTH theory builds partly on assumed competition among males, ethnic minority male applicants can be seen as a threat by male recruiters in particular. Thus, based on SMTH, one may expect to find discrimination against ethnic minority men by male recruiters in particular.

When several group identities of individuals are present at the same time, some of them are more essentialized than others, and this can result in both obstacles and opportunities for individuals (Ridgeway and Kricheli-Katz, 2013), depending on the context. Intersectional invisibility only takes place in a context where subordinate identities are regarded as non-prototypical, and, therefore, in feminine domains, that is, in female-dominated occupations, the prototypical person is a female (Purdie-Vaughns and Eibach, 2008). Given that females are prototypical in female-dominated occupations (invisibility ebbing), ethnic minority women may face even greater discrimination in these contexts as compared with occupations that are gender-neutral or male-dominated. Thus, employers’ gendered and ethnic preferences related to workers may depend on the type of occupation or job in question (e.g. Bursell, 2014), and the employer preferences are often related to ethnic stereotypes about the skills and productivity of the workers (Friberg and Midtbøen, 2018).

Empirical evidence

While there is a lot of research on gender discrimination in the labor market, the evidence of gender discrimination in recruitment seems to vary (Baert, 2018), depending on the country context and the occupational category (Birkelund et al., 2019; Di Stasio and Larsen, 2020). Some field experiments have detected gender discrimination against women in recruitment (e.g. González et al., 2019 [Spain]; Riach and Rich, 2006 [the UK]; Weichselbaumer, 2004 [Austria]). According to a recent overview of field experiments on gender discrimination since 1996, female applicants are advantaged over males in Europe, especially in female-dominated and gender-balanced occupations (Birkelund et al., 2019). Yet, some field experiments do not show any significant gender differences in callbacks in general (e.g., Albert et al., 2011 [Spain]; Baert, 2015 [Belgium]; Bygren et al., 2017 [Sweden]; Bygren and Gähler, 2021 [Sweden]).

Ethnic discrimination in recruitment, that is, discrimination against job applicants with foreign-sounding names, is well documented in research (e.g., Baert, 2018; Riach and Rich, 2002 [overviews of several countries]; Bursell, 2014 [Sweden]; Bursell et al., 2021 [Sweden]; Carlsson and Rooth, 2007 [Sweden]; Di Stasio and Larsen, 2020 [the UK, Germany, the Netherlands, Norway, and Spain]). While Carlsson and Rooth, (2007) show ethnic discrimination to be greater in more qualified jobs, according to Bursell et al. (2021) discrimination against foreign-named job applicants appears to be very similar in both high-skilled and low-skilled occupations.

Field experiments show evidence of gendered ethnic discrimination in various labor markets, that is, ethnic minority male job applicants experience more discrimination than ethnic minority female applicants (Andriessen et al., 2012⁴; Arai et al., 2016; Dahl and Krog, 2018; Liebkind et al., 2016; Midtbøen, 2016; Sidanius and Pratto, 2001). Yet, there is some inconsistency regarding this pattern (e.g. Blommaert et al., 2014), and the level and direction of gendered ethnic discrimination appear to vary by occupational gender composition (Bursell, 2014) as well as ethnic category (Di Stasio and Larsen, 2020).

Only a few field experiments examine the role of recruiter (i.e. contact person) gender in recruitment discrimination, and, to my knowledge, there is one such field experiment studying gendered ethnic discrimination. Edo et al. (2019) show that female recruiters favor female job applicants over male applicants and that French-named female and male recruiters discriminate against foreign-named males over foreign-named females, in accounting jobs in Paris and its suburbs. Also, Erlandsson (2019) shows male recruiters to contact male job applicants more often than females with identical merits, especially in gender-balanced occupations. However, some field experiments show no evidence of gender discrimination by recruiter gender in Australia (including only female-dominated occupations) (Booth and Leigh, 2010) or in Sweden, except for some indication that male recruiters favor female applicants in male-dominated occupations (significant only at the 10% level) (Carlsson, 2011). These field experiments differ from each other in several ways, especially when it comes to the number and types of occupations that are included, which can partly explain the variance in findings. Moreover, while a meta-analysis of experimental studies on employment decision-making shows nearly no general gender bias among female evaluators, it shows a pro-male bias among male evaluators, but the findings vary by the gender composition of the occupation (Koch et al., 2015).

Expectations

Based on the statistical discrimination theory, status characteristics theory, theories about homophily and in-group favoritism, as well as in line with previous research using field experiments (e.g. Baert, 2018; Di Stasio and Larsen, 2020), discrimination against ethnic minority applicants, that is, applicants with foreign-sounding names, can be assumed to take place. While previous studies on the role of the recruiter gender are few, and the results differ, theories about a (potential) gender difference in the tendency to engage in ethnic discrimination are scarce. Yet, if most female and male recruiters are native Swedes, one may assume homophily (or an in-group bias) on the basis of ethnicity to prevail among both male and female recruiters, and to result in both male and female recruiters contacting ethnic minority applicants less often than ethnic majority applicants. In addition, statistical discrimination against ethnic minority applicants can be expected to take place because of differences in average productivity, regardless of recruiter gender. Also, according to the status characteristics theory, one would expect individuals from the dominant group, that is, ethnic majority applicants, to be favored over the minority applicants by both male and female recruiters. Therefore, I will test Hypothesis 1 that there is discrimination against ethnic minority applicants over ethnic majority applicants by both female and male recruiters.

However, on the one hand, according to the double jeopardy perspective, one would theoretically expect ethnic minority women to experience more discrimination than ethnic minority men. Yet, on the other hand, based on the SMTH one can expect the opposite, that is, that ethnic minority males face more discrimination than ethnic minority females. But regarding gendered ethnic discrimination and recruiter gender, the SMTH suggests (ethnic majority) male recruiters in particular discriminate against ethnic minority men over ethnic minority women, as the theory builds partly on assumed competition among males. Also, the intersectional invisibility approach (Purdie-Vaughns and Eibach, 2008) suggests that ethnic minority women are affected by intersectional invisibility and will thereby experience less discrimination than ethnic minority men, from both male and female recruiters. In addition, one may assume gender-based homophily and in-group favoritism to prevail among female recruiters. Therefore, I will test Hypothesis 2 that there is more discrimination against ethnic minority male applicants than against ethnic minority female applicants by both female and male recruiters.

The theories outlined above suggest somewhat different, and more complex, mechanisms and outcomes regarding gendered ethnic discrimination based on the occupational type combined with recruiter gender. While homophily, and in-group favoritism, would result in discrimination against members of the out-group, regardless of the occupational type, based on the status characteristics theory and intersectional invisibility, gendered ethnic discrimination is likely to differ by the gender composition and qualification level of the occupation. Thus, rather than formulating several explicit (and contradictory) hypotheses regarding different types of occupations, an exploratory approach is applied here by examining gendered ethnic discrimination and recruiter gender in a number of occupational categories that vary in terms of qualification level and occupational gender composition.

Data and method

A correspondence test was performed in the Swedish labor market in order to study potential discrimination in recruitment. Fictitious applications were submitted to advertised vacancies, and the subsequent employer responses, that is, callbacks, were observed. The applications in this design are nearly identical within occupations, yet they differ in that they indicate the different group identities based on gender and ethnicity, that is, foreign background, as signaled by applicant names. Importantly, the method is subject to ethical concerns, mainly due to a lack of informed consent, as it is not possible to inform the employers about the study.

Experimental design

The data used for this study come from a correspondence test conducted between October 2013 and August 2020.⁵ The data used here include 5641 fictitious job applications and the employer callbacks to these applications. The jobs were primarily located in and near the three largest cities in Sweden, that is, Stockholm (including the Mälardalen area), Gothenburg, and Malmö. The vacancies were found on the largest database for job search in Sweden, that is, the Swedish Employment Agency's website. Information from the job advertisement, such as the gender of the contact person (indicated by a name), was documented when available. Vacancies that accepted applications through email or as website uploads, without requiring a social security number or filling out the applicant's qualifications in an online form, were chosen.⁶

This study only includes job applications for vacancies with information about the gender of the recruiter, that is, the contact person or employer. The gender of the recruiter is based firstly on the name of the person who contacted the (fictitious) applicant, and secondly, if there is no such information, for example, no employer response is received, it is based on the name of the contact person stated in the advertisement.⁷ According to the data collected, recruiter gender tends to be the same in the job advertisement as for the person who contacted the applicant.⁸ There are few exceptions to this (about 2% of the applications included here). This fact indicates that a potential problem with the study design does not seem to be an issue here, that is, employers do not seem to have routines where male recruiters contact male applicants and female recruiters contact female applicants. Moreover, the callback contents suggest that the person listed in the job announcement tends to be active in the recruitment decisions and is often the same person who responds to the applicant, for example, “I have read your application and would like to meet you” is an example of a typical phrase included in a callback.

The study includes vacancies in 20 occupations (which were combined into 17 analytical categories) that vary as regards gender and immigrant composition, sector, and educational qualifications (see Bursell, 2014). The occupations are also relatively representative of the Swedish labor market; applications were submitted to 14 of the 30 most common occupations, while seven of the occupations belong to the 10 most common occupations in Sweden.

One application is submitted for each vacancy by email or via a website (depending on the instructions in the job announcement). This method, in contrast to submitting more than one application to each job opening, permits the use of identical applications, rather than simply using applicants with equivalent qualifications, saves time and work for the employer (which is ethically a benefit), and limits the risk of detection. In addition, the method of using only one application avoids the risk of spillover effects (cf. Phillips, 2019). Gender, ethnicity, and some other characteristics⁹ are randomized in the applications. Gender is indicated by the name of the applicant, that is, either a distinctive female or male name. Age is held constant at 31 years in order for the applicants to be old enough to have higher education and some work experience while still not having reached too senior a position. All applicants hold occupation-specific educational qualifications and an employment background from Sweden.

Majority ethnicity, that is, a native Swede, is signaled by a set of typical Swedish names whereas minority ethnicity (indicating immigrant background) is signaled by common Arabic names and Slavic names, that is, the latter refers to typical names for individuals from the region of former Yugoslavia (presently known as Bosnia, Croatia, Kosovo, Northern Macedonia, Serbia and Slovenia). These groups represent some of the largest groups of the foreign-born population in Sweden (Statistics Sweden, 2020),¹⁰ and can be regarded as visible ethnic minority groups, meaning individuals who can by their appearance be distinguished from the ethnic majority. Employment gaps exist between individuals with these backgrounds (i.e. here indicated by Arabic- and Slavic-sounding names) and “native” Swedes, yet the gaps are generally greater for those originating from the Middle East (le Grand and Szulkin, 2002; Lundborg, 2013).

Each application includes an application letter and a simple resume, both matched to the occupation (see the Online Appendix in Bygren et al., 2017). These reveal the applicant's personal details, that is, name, date of birth, phone number, email address, and postal address. The resume presents previous jobs, education, language skills (all applicants have Swedish as a mother tongue), and computer skills. The application letter includes a biography with information about, for example, prior work experience and qualifications. Each fictitious applicant has a unique email address and a telephone number connected to a voicemail (following a few ring signals, a recording with the applicant's name and a request to leave a message is played). Here, a callback is a positive response from the employer by email or phone (automatic replies were ignored), such as a job offer, invitation to an interview, or a request for more information or to contact the recruiter, which suggests interest in the applicant.¹¹ If a callback is made, the employer is promptly told that the applicant no longer has any interest in the job.

In order to examine discrimination in different types of occupations based on occupational gender composition (Statistics Sweden, 2012), three categories are created for the analysis. Occupations in which the proportion of women is higher than 60% are categorized as female-dominated occupations, that is, store personnel/cashier, financial assistant, preschool teacher, elementary school teacher, cleaner, accountant, receptionist, nurse, and assistant nurse. Occupations in which the proportion of men is higher than 60% are categorized as male-dominated, and these include carpenter, driver, salesperson, computer science engineer/programmer, and engineer in machine technology/industrial economics/electronics. The remaining occupations, that is, high school teacher, chef, and auditor, in which the proportion of women is 40% to 60% are categorized as gender-balanced occupations. Furthermore, the following occupations are defined as highly qualified occupations in the analysis: accountant, auditor, computer specialist, engineer, nurse, preschool teacher as well as elementary and high school teacher.¹²

Results

The sample has a total callback rate of 36.7%. This means that 36.7% of the submitted job applications resulted in a callback, that is, a positive response from the employer. The callback rate is calculated by dividing the number of positive callbacks by the total number of applications sent.

Table 1 presents the number of job applications, the proportion of recruiters and job applicants, and the callback rates for the different categories. In line with most previous studies on ethnic discrimination in recruitment (cf. Baert, 2018), job applicants with foreign-sounding names receive fewer callbacks than applicants with Swedish-sounding names, the callback rates being 44.0% for Swedish-named applicants and 29.4% for foreign-named applicants, yielding a difference of 14.6 percentage points. This difference is statistically significant (p < 0.001) and can be taken as evidence of ethnic discrimination in the Swedish labor market.

Table 1.

Callback rates by job applicant's ethnicity (i.e. a foreign or a Swedish name) and gender, and recruiter's gender.

	Callback rate				Ratio	Difference
	Number of jobs	All	Swedish name (N = 3233)	Foreign name (N = 2408)	Swedish versus foreign	Swedish- foreign
Aggregated results	5641	36.67	43.98	29.36	1.50***	14.6***
Woman	2857	37.86	43.71	32.00	1.37***	11.7***
Man	2784	35.50	44.27	26.72	1.66***	17.5***
Female recruiter	2682	38.87	46.76	30.97	1.51***	15.8***
Female recruiter × woman	1315	41.93	48.72	35.14	1.39***	13.6***
Female recruiter × man	1367	35.95	44.83	27.08	1.66***	17.7***
Male recruiter	2959	34.70	41.59	27.80	1.50***	13.8***
Male recruiter × woman	1542	34.39	39.63	29.16	1.36***	10.5***
Male recruiter × man	1417	35.05	43.76	26.35	1.66***	17.4***

*p < 0.05, **p < 0.01, ***p < 0.001.

Both male and female recruiters engage in ethnic discrimination by contacting Swedish-named applicants more often than applicants with foreign-sounding names, and these differences are statistically significant (p < 0.001). Thus, Hypothesis 1, that both female and male recruiters discriminate against ethnic minority applicants over ethnic majority applicants, is supported. The callback ratios that are presented in Table 1 indicate that there appears practically no overall difference in the level or direction of ethnic discrimination between male recruiters (a ratio of 1.50) and female recruiters (a ratio of 1.51).

In addition to this, there are clear patterns of different levels of ethnic discrimination toward male and female job applicants. For male applicants, the callback ratio is 1.66 overall, compared to 1.37 for female applicants. In other words, for every 10 job applications sent by a Swedish-named male applicant, a foreign-named male applicant would have to send close to 17 applications to achieve a similar number of callbacks. For female candidates, the difference is smaller: for every 10 job applications sent by a Swedish-named female applicant, a foreign-named female applicant would have to send close to 14 applications to achieve a similar expected number of callbacks. Thus, I find patterns of gendered ethnic discrimination consistent with the previous literature (e.g. Arai et al., 2016; Dahl and Krog, 2018; Liebkind et al., 2016; Midtbøen, 2016), and support for Hypothesis 2. Finally, these patterns of gendered ethnic discrimination do not seem to depend on recruiter gender.

Figure 1 visualizes the callback rates by applicant gender and recruiter gender, which are displayed in Table 1, but confidence intervals are also shown. The figure portrays the gendered pattern in ethnic discrimination in line with the SMTH and the intersectional invisibility approach: the ethnic gap in callbacks is somewhat larger among men than women. Also, foreign-named men have the lowest callback rate regardless of recruiter gender. Moreover, Figure 1 illustrates a slight own-gender bias in callback rates for Swedish-named applicants, yet the difference is statistically significant only at a 10% level for male recruiters and it does not reach statistical significance for female recruiters.

Figure 1.

Callback rates by recruiter gender and job applicant gender and ethnicity (i.e. a foreign or a Swedish name). Ninety-five percent confidence intervals.

Theoretically, the occupational context is expected to be important for discrimination outcomes. While there are too few observations for certain occupations to allow for a meaningful analysis of occupation-specific discrimination patterns, I run separate analyses for female and male recruiters based on the qualification level of the occupation, that is, whether post-secondary education tends to be required in the occupation or not. Figure 2 visualizes the callback rates for male and female recruiters separately by the qualification level of the occupation. In Figure 2, the ethnic gaps appear largest among men in highly qualified occupations.

Figure 2.

Callback rates for combinations of recruiter gender and job applicant gender and ethnicity (i.e. a foreign or a Swedish name), by the qualification level of occupation. Ninety-five percent confidence intervals.

Figure 3 illustrates the callback rates for female and male recruiters separately by the occupational gender composition. Figure 3 shows clear variation in the ethnic gaps in callbacks among men and women, respectively, by recruiter gender and the occupational gender composition.

Figure 3.

Callback rates for combinations of recruiter gender and job applicant gender and ethnicity (i.e., a foreign or a Swedish-sounding name), by the gender composition of occupation. Ninety-five percent confidence intervals.

To further assess the gendered differences in ethnic discrimination, that is, whether there is ethnic discrimination and, if so, does it differ between male and female applicants, as well as to control whether the results are due to sampling variation in occupation and other job-related characteristics (e.g. a time period when the job application was submitted and whether an external recruitment or personnel agency was involved), I run a linear probability regression (Table 2). The models include a variable “Foreign name” to measure ethnic discrimination, a variable “Man” to measure gender discrimination, and an interaction term “Foreign name*Man” to measure gendered ethnic discrimination. Model 1 reconfirms Hypothesis 1 regarding ethnic discrimination that was discussed above. Also, Model 1 is in line with Hypothesis 2, that there is more (ethnic) discrimination against men with foreign-sounding names than women with foreign-sounding names by both male and female recruiters.

Table 2.

Regression of callbacks (dichotomous) on the gender of recruiter, ethnicity (i.e. a Swedish or a foreign name), and gender of job applicant.

		All		Highly qualified occupations		Less qualified occupations
	Model 1 all	Model 2 female recruiters	Model 3 male recruiters	Model 4 female recruiters	Model 5 male recruiters	Model 6 female recruiters	Model 7 male recruiters
Foreign name	−0.106***	−0.126***	−0.088***	−0.102*	−0.174***	−0.146***	−0.055*
	(0.018)	(0.026)	(0.024)	(0.040)	(0.051)	(0.035)	(0.027)
Man	0.002	−0.031	0.028	−0.001	−0.014	−0.055	0.046
	(0.016)	(0.024)	(0.022)	(0.036)	(0.041)	(0.033)	(0.027)
Foreign name × man	−0.059*	−0.051	−0.060	−0.103*	0.001	−0.001	−0.084*
	(0.024)	(0.035)	(0.033)	(0.052)	(0.068)	(0.046)	(0.038)
Constant	0.477***	0.527***	0.434***	0.661***	0.644***	0.378***	0.333***
	(0.015)	(0.022)	(0.020)	(0.029)	(0.033)	(0.033)	(0.026)
R-squared	0.130	0.147	0.112	0.095	0.076	0.067	0.056
N	5641	2682	2959	1328	870	1354	2089
	Female-dominated occupations		Male-dominated occupations		Gender-balanced occupations
	Model 8 female recruiters	Model 9 male recruiters	Model 10 female recruiters	Model 11 male recruiters	Model 12 female recruiters	Model 13 male recruiters
Foreign name	−0.119***	−0.171***	−0.144*	−0.025	−0.131	−0.119*
	(0.031)	(0.037)	(0.058)	(0.036)	(0.093)	(0.057)
Man	−0.041	−0.010	−0.013	0.039	−0.004	0.068
	(0.028)	(0.035)	(0.056)	(0.034)	(0.086)	(0.055)
Foreign name × man	−0.044	0.031	−0.063	−0.125*	−0.096	−0.045
	(0.041)	(0.052)	(0.079)	(0.050)	(0.125)	(0.081)
Constant	0.536***	0.400***	0.474***	0.462***	0.484***	0.407***
	(0.024)	(0.029)	(0.057)	(0.035)	(0.078)	(0.044)
R-squared	0.175	0.217	0.091	0.061	0.051	0.044
N	1864	987	588	1436	230	536

*p < 0.05, **p < 0.01, ***p < 0.001.

Note: Separate models by recruiter gender, occupational gender composition, and the qualification level of the occupation. Linear probability model estimates with robust standard errors in parentheses. The coefficients in all models are conditional on the following (indicator) control variables: occupation, time period, and external recruiter. All estimated coefficients are included in Appendix A.

Yet, in Models 2 and 3, the estimation regarding the interaction effect of gender and foreign name on applicants’ callbacks, among male and female recruiters, respectively, does not reach statistical significance at the 5% level here (possibly due to a smaller sample size), that is, the precision of the estimate decreases considerably while the size of the effect remains when controls are added in the model. However, it can be noted that in Model 3 the coefficient for “Foreign name*Man” for male recruiters is statistically significant at the 10% level. Thus, the effect of recruiter gender on gendered ethnic discrimination seems to be somewhat moderated by other occupational, and job-related, factors.

Models 4 and 7 show that, when controls are included, foreign-named men appear to experience statistically significant gendered ethnic discrimination not only by female recruiters in highly qualified occupations but also by male recruiters in less qualified occupations. When it comes to gendered ethnic discrimination in occupational categories based on the occupational gender composition (Model 8 to 13), the coefficient for the interaction of foreign names and applicant gender is statistically significant only in one model. This is the case in Model 11, confirming a clear negative and statistically significant effect of being a foreign-named man on the callbacks among male recruiters in male-dominated occupations.

Conclusion

The results presented here show, in line with most of the previous research from Europe (e.g. Arai et al., 2016; Dahl and Krog, 2018; Liebkind et al., 2016; Midtbøen, 2016), that foreign-named male job applicants face more recruitment discrimination than foreign-named female applicants. Also, recruiter gender does not seem to matter for the direction of ethnic discrimination. Both male and female recruiters discriminate against job applicants with a foreign-sounding name over Swedish-named applicants, and this tendency is stronger for male applicants.

Moreover, female recruiters discriminate against foreign-named male applicants over foreign-named female applicants especially in highly qualified occupations, whereas male recruiters discriminate against foreign-named males over foreign-named females in less qualified occupations and male-dominated occupations in particular. These findings highlight the importance of intersectionality, and how the group characteristics, or group identities based on gender and ethnicity (i.e. a foreign-sounding name), and the combinations of these, form a base for differential treatment of job applicants, both by male and female recruiters in different occupational contexts.

While there is no support for the multiple burden theory that ethnic minority women, in general, would face more discrimination than ethnic minority men, the foreign-named men seem to be subordinate male targets although this varies by the occupational category in combination with the gender of the recruiter. Thus, the results lend some support to the theoretical approach that having the double subordinate status of a female minority can, through intersectional invisibility, protect foreign-named women from the strongest discrimination, while the occupational context matters. At the same time, the findings suggest that status characteristics related to ethnic minority males together with occupational context play a role in discrimination outcomes among female and male recruiters.

The results presented here show, in line with the SMTH and most previous research, that ethnic minority men in particular are subject to discrimination. This appears to be a pattern, although the explorative approach indicates gendered ethnic discrimination against foreign-named men to be stronger in male-dominated and less qualified occupations by male recruiters and in highly qualified occupations by female recruiters. This discrimination pattern calls for future research to address the mechanisms behind the discrimination against foreign-named men.

There are limitations to this study. One is that discrimination may also take place later in the recruitment process, that is, in the final hiring decisions which are not studied here. Almost 90% of the discrimination against ethnic minority applicants appears to occur at this early stage, rather than later in the recruitment process (Allasino et al., 2004; Attström, 2007; Cediey and Foroni, 2008). However, in a recent meta-study, Quillian et al. (2020) argue that this is due to most applicants of any race or ethnicity being dropped at the first stage where callback rates are typically low—thereby resulting in few applicants at risk of discrimination at the later stage—and show ethnic and racial discrimination to be greater later at the job offer stage than in callbacks at the early stage. Thus, while the method used here is superior to other methods in detecting labor market discrimination, it may not accurately measure the extent of discrimination in the labor market at large. Also, in Sweden, permanent workers generally have a relatively high employment protection, that is, stronger than the OECD average (OECD, 2021), which is likely to increase the pressure on recruiters in their recruitment decisions, in comparison to other contexts.

Moreover, one cannot know with certainty who makes the recruitment decisions. Although, based on the content of the employer responses, it is likely that the person who contacts the applicant, or is mentioned in the job advertisement, is involved in the hiring decision. Moreover, because of gender segregation in the labor market, the proportion of female and male recruiters is likely to differ by the type of occupation. As shown here, the patterns for gendered ethnic discrimination among male and female recruiters differ depending on the occupational gender composition and the qualification level of the occupation. Thus, the discrimination patterns are also likely to vary between specific occupations. Because of data limitations, separate occupations cannot be studied here in a meaningful way. Yet, it would be interesting to study occupation-specific discrimination patterns on this topic in future research.

To conclude, the findings presented here show evidence of widespread discrimination against foreign-named men in the Swedish labor market. The results suggest that while gender diversity among recruiters appears not to be crucial for gendered ethnic discrimination in general, it plays a varying role in recruitment outcomes, that is, who gets a chance to proceed in the recruitment process, depending on the occupational context. Finally, the results hint that female recruiters may contribute to the vertical ethnic segregation of foreign-named men by discriminating against them in highly qualified occupations.

Footnotes

Acknowledgements

Valuable comments from Magnus Bygren,Michael Gähler,Lynn Prince Cooke,Karin Halldén,Edvard Nergård Larsen,Charlotta Magnusson,and Magnus Nermo as well as three anonymous reviewers are greatly appreciated.

Funding

The author disclosed receipt of the following financial support for the research,authorship,and/or publication of this article: This research was supported by the Swedish Research Council for Health,Working Life and Welfare (Forte grant 2012-0587);and by the Strategic Research Council (SRC,established within the Academy of Finland),FLUX consortium (decision number 345130).

ORCID iD

Anni Erlandsson

Author biography

Anni Erlandsson is a Senior Researcher at the INVEST Research Flagship Centre,University of Turku,Finland,and affiliated with the Department of Sociology,the Demography Unit,Stockholm University. She has previously worked at the Swedish Institute for Social Research (SOFI),Stockholm University. Her main research interests include family sociology and labor market inequality. Her work appears in European Sociological Review,Work and Occupations,and Population Research and Policy Review .

References

Albert

Escot

Fernández-Cornejo

(2011) A field experiment to study sex and age discrimination in the Madrid labour market. The International Journal of Human Resource Management 22(2): 351–375.

Allasino

Reyneri

Venturini

, et al. (2004) Labour Market Discrimination Against Migrant Workers in Italy. International Migrations Papers 67. Geneva: International Labour Office.

Andriessen

Nievers

Dagevos

, et al. (2012) Ethnic discrimination in the Dutch labor market: Its relationship with job characteristics and multiple group membership. Work and Occupations 39(3): 237–269.

Arai

Bursell

Nekby

(2016) The reverse gender gap in ethnic discrimination: Employer stereotypes of men and women with Arabic names. International Migration Review 50(2): 385–412.

Arrow

(1973) The theory of discrimination. In: Ashenfelter

Rees

(eds) Discrimination in Labor Markets. Princeton, NJ: Princeton University Press, 3–42.

Attström

(2007) Discrimination against Native Swedes of Immigrant Origin in Access to Employment. International Migrations Papers 86E. Geneva: International Labour Office.

Baert

(2015) Field experimental evidence on gender discrimination in hiring: Biased as Heckman and Siegelman predicted? Economics: The Open-Access, Open-Assessment E-Journal 9(25): 1–11.

Baert

(2018) Hiring discrimination: An overview of (almost) all correspondence experiments since 2005. In: Gaddis

(ed.) Audit Studies: Behind the Scenes with Theory, Method, and Nuance. Cham, Switzerland: Springer, 63–77.

Becker

(1971) The Economics of Discrimination. Chicago, IL: University of Chicago Press.

10.

Berdahl

Moore

(2006) Workplace harassment: Double jeopardy for minority women. Journal of Applied Psychology 91: 426–436.

11.

Bielby

Baron

(1986) Men and women at work: Sex segregation and statistical discrimination. American Journal of Sociology 91: 759–799.

12.

Birkelund

Janz

Larsen

(2019) Do males experience hiring discrimination in female-dominated occupations? An overview of field experiments since 1996. GEMM working paper. Available at: https://gemm2020.eu/wp-content/uploads/2018/12/Gender-Discrimination-a-summary-1.pdf

13.

Blau

(1984) Occupational segregation and labor market discrimination. In: Reskin

(ed.) Sex Segregation in the Workplace: Trends, Explanations, Remedies. Washington, D.C: National Academy Press, 117–143.

14.

Blau

Kahn

(2017) The gender wage gap: Extent, trends, and explanations. Journal of Economic Literature 55(3): 789–865.

15.

Blommaert

Coenders

van Tubergen

(2014) Discrimination of Arabic-named applicants in The Netherlands: An internet-based field experiment examining different phases in online recruitment procedures. Social Forces 92: 957–982.

16.

Booth

Leigh

(2010) Do employers discriminate by gender? A field experiment in female-dominated occupations. Economics Letters 107(2): 236–238.

17.

Brewer

Kramer

(1985) The psychology of intergroup attitudes and behavior. Annual Review of Psychology 36(1): 219–243.

18.

Bursell

(2014) The multiple burdens of foreign-named men—evidence from a field experiment on gendered ethnic hiring discrimination in Sweden. European Sociological Review 30(3): 399–409.

19.

Bursell

Bygren

Gähler

(2021) Does employer discrimination contribute to the subordinate labor market inclusion of individuals of a foreign background? Social Science Research 98: 102582.

20.

Bygren

Erlandsson

Gähler

(2017) Do employers prefer fathers? Evidence from a field experiment testing the gender by parenthood interaction effect on callbacks to job applications. European Sociological Review 33(3): 337–348.

21.

Bygren

Gähler

(2021) Are women discriminated against in countries with extensive family policies? A piece of the ‘welfare state paradox’ puzzle from Sweden. Social Politics: International Studies in Gender, State & Society 28(4): 921–947.

22.

Carlsson

(2011) Does hiring discrimination cause gender segregation in the Swedish labor market? Feminist Economics 17: 71–102.

23.

Carlsson

Rooth

D-O

(2007) Evidence of ethnic discrimination in the Swedish labor market using experimental data. Labour Economics 14: 716–729.

24.

Cediey

Foroni

(2008) Discrimination in Access to Employment on Grounds of Foreign Origin in France. International migration papers 85E. Geneva: International Labour Office.

25.

Charles

(2011) A world of difference: International trends in women's economic status. Annual Review of Sociology 37: 355–371.

26.

Correll

Benard

(2006) Biased estimators? Comparing status and statistical theories of gender discrimination. Advances in Group Processes 23: 89–116.

27.

Dahl

Krog

(2018) Experimental evidence of discrimination in the labour market: Intersections between ethnicity, gender, and socio-economic status. European Sociological Review 34(4): 402–417.

28.

Di Stasio

Larsen

(2020) The racialized and gendered workplace: Applying an intersectional lens to a field experiment on hiring discrimination in five European labor markets. Social Psychology Quarterly 83(3): 229–250.

29.

Edlund

Öun

(2016) Who should work and who should care? Attitudes towards the desirable division of labour between mothers and fathers in five European countries. Acta Sociologica 59(2): 151–169.

30.

Edo

Jacquemet

Yannelis

(2019) Language skills and homophilous hiring discrimination: Evidence from gender and racially differentiated applications. Review of Economics of the Household 17(1): 349–376.

31.

England

(1992) Comparable Worth: Theories and Evidence. Hawthorne, NY: Aldine de Gruyter.

32.

Erlandsson

(2019) Do men favor men in recruitment? A field experiment in the Swedish labor market. Work and Occupations 46(3): 239–264.

33.

European Institute for Gender Equality (2017) Gender Equality Index 2017. Measuring Gender Equality in the European Union 2005–2015. Vilnius: European Institute for Gender Equality.

34.

Friberg

Midtbøen

(2018) Ethnicity as skill: Immigrant employment hierarchies in Norwegian low-wage labour markets. Journal of Ethnic and Migration Studies 44(9): 1463–1478.

35.

Gaddis

(2018) An introduction to audit studies in the social sciences. In: Gaddis

(ed.) Audit Studies: Behind the Scenes with Theory, Method, and Nuance. Cham, Switzerland: Springer, 3–44.

36.

Goldin

Rouse

(2000) Orchestrating impartiality: The impact of “blind” auditions on female musicians. American Economic Review 90(4): 715–741.

37.

González

Cortina

Rodríguez

(2019) The role of gender stereotypes in hiring: A field experiment. European Sociological Review 35(2): 187–204.

38.

Grönlund

Halldén

Magnusson

(2017) A Scandinavian success story? Women’s labour market outcomes in Denmark, Finland, Norway and Sweden. Acta Sociologica 60(2): 97–119.

39.

Hewstone

Rubin

Willis

(2002) Intergroup bias. Annual Review of Psychology 53(1): 575–604.

40.

Kandel

(1978) Homophily, selection, and socialization in adolescent friendships. American Journal of Sociology 84(2): 427–436.

41.

King

(1988) Multiple jeopardy, multiple consciousness: The context of a black feminist ideology. Signs: Journal of Women in Culture and Society 14(1): 42–72.

42.

Koch

D'Mello

Sackett

(2015) A meta-analysis of gender stereotypes and bias in experimental simulations of employment decision making. Journal of Applied Psychology 100(1): 128.

43.

Le Grand

Szulkin

(2002) Permanent disadvantage or gradual integration: Explaining the immigrant–native earnings gap in Sweden. Labour (Committee on Canadian Labour History) 16(1): 37–64.

44.

Liebkind

Larja

Brylka

(2016) Ethnic and gender discrimination in recruitment: Experimental evidence from Finland. Journal of Social and Political Psychology 4: 403–426.

45.

Lippens

Vermeiren

Baert

(2023) The state of hiring discrimination: A meta-analysis of (almost) all recent correspondence experiments. European Economic Review 151: 104315.

46.

Lundborg

(2013) Refugees’ employment integration in Sweden: Cultural distance and labor market performance. Review of International Economics 21(2): 219–232.

47.

Mandel

Semyonov

(2006) A welfare state paradox: State interventions and women's employment opportunities in 22 countries. American Journal of Sociology 111: 1910–1949.

48.

McPherson

Smith-Lovin

Cook

(2001) Birds of a feather: Homophily in social networks. Annual Review of Sociology 27(1): 415–444.

49.

Midtbøen

(2016) Discrimination of the second generation: Evidence from a field experiment in Norway. Journal of International Migration and Integration 17: 253–272.

50.

Nelson

Probst

(2004) Multiple minority individuals: Multiplying the risk of workplace harassment and discrimination. In: Chin

(ed.) The Psychology of Prejudice and Discrimination: Ethnicity and Multiracial Identity. Westport, CT: Praeger, 193–217.

51.

OECD (2014) Finding the Way: A Discussion of the Swedish Migrant Integration System. Paris: OECD publishing.

52.

OECD (2017) OECD Employment Outlook 2017. Paris, France: OECD Publishing.

53.

OECD (2021) OECD Economic Surveys: Sweden 2021. Paris: OECD Publishing.

54.

Phelps

(1972) The statistical theory of racism and sexism. American Economic Review 62(4): 659–661.

55.

Phillips

(2019) Do comparisons of fictional applicants measure discrimination when search externalities are present? Evidence from existing experiments. The Economic Journal 129(621): 2240–2264.

56.

Purdie-Vaughns

Eibach

(2008) Intersectional invisibility: The distinctive advantages and disadvantages of multiple subordinate-group identities. Sex Roles 59(5–6): 377–391.

57.

Quillian

Lee

Oliver

(2020) Evidence from field experiments in hiring shows substantial additional racial discrimination after the callback. Social Forces 99(2): 732–759.

58.

Riach

Rich

(2002) Field experiments of discrimination in the market place. The Economic Journal 112(483): F480–F518.

59.

Riach

Rich

(2006) An experimental investigation of sexual discrimination in hiring in the English labor market. Advances in Economic Analysis and Policy 5(2): 1–20.

60.

Ridgeway

(2011) Framed by Gender: How Gender Inequality Persists in the Modern World. New York: Oxford University Press.

61.

Ridgeway

Smith-Lovin

(1999) The gender system and interaction. Annual Review of Sociology 25: 191–216.

62.

Ridgeway

Correll

(2004) Unpacking the gender system: A theoretical perspective on gender beliefs and social relations. Gender & Society 18(4): 510–531.

63.

Ridgeway

Kricheli-Katz

(2013) Intersecting cultural beliefs in social relations: Gender, race, and class binds and freedoms. Gender & Society 27(3): 294–318.

64.

Sandberg

(2018) Competing identities: A field study of in-group bias among professional evaluators. The Economic Journal 128(613): 2131–2159.

65.

Sidanius

Pratto

(2001) Social Dominance: An Intergroup Theory of Social Hierarchy and Oppression. Cambridge, UK: Cambridge University Press.

66.

Sidanius

Veniegas

(2000) Gender and race discrimination: The interactive nature of disadvantage. In: Oskamp

(ed.) Reducing Prejudice and Discrimination. The Claremont Symposium on Applied Social Psychology. Mahwah, NJ: Lawrence Erlbaum Associates, 47–69.

67.

Statistics Sweden (2012) Yrkesstrukturen i Sverige 2012 [The occupational structure in Sweden 2012]. Statistiskt meddelande AM 33 SM 1401. SCB.

68.

Statistics Sweden (2020) Utrikes födda i Sverige [The foreign-born in Sweden]. Available at: https://www.scb.se/hitta-statistik/sverige-i-siffror/manniskorna-i-sverige/utrikes-fodda (accessed 1 July 2021).

69.

Tajfel

Billig

Bundy

, et al. (1971) Social categorization and intergroup behaviour. European Journal of Social Psychology 1(2): 149–178.

70.

United Nations Development Programme (2016) Human Development Report 2016. Human Development for Everyone. New York, NY: United Nations Development Programme.

71.

Weichselbaumer

(2004) Is it sex or personality? The impact of sex stereotypes on discrimination in applicant selection. Eastern Economic Journal 30(2): 159–186.