Is employer collection of diversity data attractive to potential job seekers? Ethnicity and sex differences and a UK–Sweden comparison

Abstract

Purpose

Many organisations monitor statistics on the background of job applicants to inform diversity management, a practice known as equality monitoring (EM). The study examines perceptions of EM and employers that use it. Additionally, it aims to assess potentially salient group differences in attitudes towards EM, focussing on perceived history of employment discrimination, ethnicity, sex, and a comparison between the UK and Sweden – two countries which differ extensively in EM prevalence.

Design/methodology/approach

A cross-sectional self-report survey assessed attitudes toward EM, attraction to employers using it, pro-equality and diversity attitudes, perceived history of employment discrimination and background characteristics (e.g. ethnicity and sex), and compared a UK and Sweden sample (N = 925).

Findings

The results reveal positive perceptions of EM overall. Although no differences were observed between UK ethnic majority and minority respondents, White British men rate employers using EM as less attractive with increasing levels of perceived past discrimination. Women have more positive perceptions than men. Finally, the UK sample rated EM more positively than the Sweden sample.

Originality/value

Despite EM being widespread, the study is the first to investigate detailed perceptions of it, making group and country comparisons. Results support the use of EM in HRM but highlight the need for clear communication to avoid confusion with positive discrimination, which is perceived negatively in some majority group members, and to allay fears of data misuse. Recommendations are made for future implementation.

Citation

Heath, A.J., Carlsson, M. and Agerström, J. (2023), "Is employer collection of diversity data attractive to potential job seekers? Ethnicity and sex differences and a UK–Sweden comparison", Personnel Review, Vol. 52 No. 7, pp. 1900-1915. https://doi.org/10.1108/PR-10-2021-0735

Introduction

Statistical evidence can play a vital role in uncovering areas of inequality, disadvantage and discrimination, especially subtler indirect discrimination. For organisations, collection of statistics on personal characteristics of employees and job applicants can permit comparison with external benchmarks, such as the proportion of groups in the working age population, and can be utilised to inform and evaluate policies aimed at increasing equality in recruitment, retention, and promotions. Transparency of this data can boost public awareness and provoke wider discussion and sensitivity to issues of inclusivity and discrimination (Makkonen, 2016).

Equality monitoring (EM) is a term commonly used in the UK to describe the processes surrounding collection of equality data. We define it as the collection and use of personal data, usually on legally protected characteristics (for e.g. ethnicity, sex etc.), to promote inclusion, reduce inequality and increase diversity reflective of wider society. We suggest that EM differs from mere equality data collection in requiring actions to be taken based on such data, hence the emphasis on use of data. EM is achieved often through the administration of anonymous and voluntary self-declaration forms which ask about personal characteristics such as age and gender. The practice differs in prevalence from country to country (Huddleston, 2017). It is especially widespread in large and public sector organisations in countries such as the UK and US and often used to monitor data on job applicants during the recruitment phase. Despite being common, it appears to be scant research on public perceptions of equality data collection in general, aside from some EU survey questions, and, to our knowledge, there is none that looks at perceptions of equality monitoring as an HR diversity management practice. Do such diversity initiatives attract or deter potential job applicants and what do majority groups and the minority groups, whom such data collection is intended to help, think about them?

The current study examines public perceptions of equality monitoring in recruitment, and their rated attraction to organisations that use it. We use a cross-sectional survey to measure perceptions in majority and minority ethnic groups, men and women, and provide a country comparison between the nation that does most EM, the UK, and one of the countries doing the least, Sweden. Since there can be considerable benefits in collecting this data for organisations to monitor and evaluate their diversity management and to provide statistics on under-representation in society, it is important to examine if this is perceived favourably by the public, and that potential job applicants especially are comfortable in completing monitoring forms. The results, if positive, might promote a wider uptake of EM across the private sector, and in countries that currently do little of such data collection and monitoring and where the full extent of under-representation in the labour market and society goes unrecorded.

Equality monitoring – two ends of the spectrum: the UK and Sweden

The United Kingdom leads on EM in Europe (Farkas, 2017; Huddleston, 2017) where it has become commonplace in the public sector, and frequently in the private sector, for job applicants, employees or service users to be asked to complete voluntary self-declaration forms. Often the stated reason provided for the data collection is to comply with obligations under the law (especially with public employers), however, other stated reasons relate to internal monitoring of diversity and inclusion. In the UK, since the Equality Act 2010, employers are legally obliged to collect statistics (for example, on the gender pay gap), and public bodies must set targets based on them under the Public Sector Equality Duty (2011). Although this began as an exercise in compliance with legal obligations and in response to pressure from NGOs for data on discrimination, it is increasingly being used as a method of evidence-based diversity management. This manifests in recruitment where the data from the self-declaration forms are used for internal monitoring and policy guidance. In contrast, in Sweden, equal opportunities legislation does not compel private employers or public bodies to produce equality statistics or act on them. Equality data is therefore not systematically used for monitoring, policymaking, planning or in discrimination cases (Huddleston, 2017), and here HR diversity management practices are less systematically evidence-based.

Public perceptions of EM

Within the general population, several EU Special Eurobarometer surveys have found a broad degree of support for workplace monitoring to combat discrimination (for example, EU Special Barometer, 2008). In 2015, support for workplace diversity measures in a UK sample was the highest in the EU. Support for collecting data on recruitment practices and workforce composition specifically was 92% and 85% respectively, while in Sweden the figures were lower though still in favour at 86% and 68% (Eurobarometer, 2015). Apart from these Eurobarometer questions on monitoring, perceptions of EM seem not to have been researched and we can find no published research that has studied perceptions of the process of data collection and monitoring specifically by employers or how people feel about filling in monitoring forms.

Barriers to EM

Equality data monitoring could be conflated with strong affirmative action or positive discrimination, such as quotas or preferential hiring of minorities, and which may conflict with deeply held values of fairness and meritocracy. Some members of the public may mistrust even anonymised personal data collection and have concerns about what it is used for, with a potential conflict between protection of disadvantaged groups and right to privacy. In Sweden, for example, there is an association in the public consciousness with recording race or ethnicity in crime statistics and concerns it could fuel anti-immigrant sentiment (Wikström and Hübinette, 2021). In practical terms, the collection of certain types of data might be problematic because universally agreed definitions, for example of ethnic background, are difficult, especially in Sweden where there is no national census and, if ethnicity is recorded at all, it follows the broad categorisation of Swedish- versus foreign-born origin. There has also sometimes been a misconception that data protection laws prohibit the collection of sensitive data (Al Zubaidi, 2012), while GDPR in the EU clarifies the legality of anonymous and pseudo-anonymous data collection.

Strong recommendations to pursue EM have been made by regional bodies like the EU and prominent NGOs (Makkonen, 2016). Given the potential advantages of EM, research on perceptions can uncover where any public concerns lie and be used to support a case for wider implementation in countries where it is less prevalent and increased use in HRM in countries where it already exists.

Responses to diversity initiatives in recruitment: theoretical underpinnings

While the public might be broadly positive about anti-discrimination employment monitoring, there is reason to believe that majority and under-represented groups may respond differently given research on reactions to diversity management practices more generally. Recruitment materials, websites, job advertisements, and word of mouth accounts are typical aspects that provide signals which job applicants use to evaluate a potential employer (Rynes, 1991). Social identity approaches predict that individual and group identity will selectively motivate this information processing (Ashforth and Mael, 1989), and personal and ethical values will also guide these judgements (for example, social responsibility values, Belinda et al., 2018). For members of under-represented groups, who may have been subjected to past discrimination and anticipate the possibility in the future, group identity will be especially salient in motivating information processing about employers. This results in a bias to look for inclusivity in recruitment literature or warning signals of the opposite (such as absence of such information). Complimentary to the signalling and social identity approaches is the theory of person-organisation fit, based on Schneider’s (1987) attrition-selection-attrition model. This describes how job seekers are attracted to organisations they perceive share their values, beliefs, personality and culture. High perceptions of fit with personal values and characteristics will result in higher attraction and subsequent identification (Walker and Hinojosa, 2014).

These theoretical approaches help explain observed differences in the research on majority and underrepresented groups. Thus, women and racial minorities are more supportive of workplace diversity management initiatives (Soni, 2000; Strauss et al., 2003). Signals of inclusivity in recruitment materials, such as equality and diversity statements (Avery, 2003; Flory et al., 2021; Ng and Burke, 2005), affirmative action statements (Cropanzano et al., 2005; Highhouse et al., 1999), and multicultural images or representation (Perkins et al., 2000; Pietri et al., 2018), have all been found to have positive impacts on outcomes such as organisational attraction, and intentions to pursue a job. Diversity information in recruitment, and diversity initiatives in general, sometimes have positive effects on majority group organisational attraction (Flory et al., 2021; James et al., 2001) and sometimes negative effects (Kravitz, 2008; Plaut et al., 2011). Individual differences such as prejudice, intolerance (Brown et al., 2006) and personality (Strauss et al., 2003) may factor into this. Equal employment opportunities (EEO) initiatives, emphasising equal treatment to everyone, are more positively received by majority group members than affirmative action (AA) that implies preferential treatment (Walker and Hinojosa, 2014). EEO is identity-blind as it promotes equality for everyone, regardless of group identity, whereas AA type statements are identity-conscious, naming groups for employment or special treatment. The latter may cause feelings of threat and exclusion in majority group members by the emphasis on named minorities (Parker et al., 1997).

There might also be occasions when underrepresented groups are put off by diversity initiatives and recruitment statements. Not all studies have found positive effects, possibly though fears of tokenism (Leibbrandt and List, 2018). Being hired simply because of one’s group status into an environment in which one is a small minority would not be appealing and may well be a deleterious experience. Women, often a disadvantaged group albeit not a minority group, sometimes respond more positively to EEO over AA, while racial minorities tend to respond better to AA in some literature (Wilton et al., 2015). Because the point of EM is to collect and monitor data on protected characteristics, it is therefore inherently identity-conscious rather than identity blind, and could trigger fears of tokenism, so that differences in group status could go either way. This leads to our first two research questions:

RQ1.

Do UK minority and majority ethnic groups differ in their attitudes towards EM?

RQ2.

Are there sex differences in attitudes towards EM in the UK sample?

The social identity approach predicts that the more an identity category is activated, the more it is available to guide cognition. It follows that under-represented group members who perceive they have been the subject of discrimination in the past might be especially guided by their group identity in making judgements of potential employers, and in their attitudes towards diversity initiatives like equality monitoring. This leads to our next research question.

RQ3.

Does history of discrimination interact with ethnicity to effect attraction to organisations using equality monitoring?

The UK and Sweden represent different ends of the spectrum in the use of EM. Previous European surveys suggest high support for the general principle of workplace monitoring, including data from Swedish samples where it is nonetheless infrequent. However, in practice, the idea of filling in EM forms may be alien to a Swedish sample who could well be wary of providing sensitive personal data when applying for job vacancies and sceptical about its use. The comparison between the UK and Sweden, being almost two ends of a spectrum in experience of EM, leads us to pose the question of how these populations differ in attitudes towards EM and employer use of it.

RQ4.

How do UK and Swedish samples differ in their attitudes towards EM?

Method

Participants

In total 925 participants were included in the study across both countries [1], and anonymous data were obtained using the Prolific.co research participation platform during March and April 2021. Payment to participants was £1.25.

The UK sample consisted of 741 participants, and we pre-defined our target population using Prolific filters for UK nationality and UK country of residence, with approximately 50–50 split of sex and ethnic group using White British (majority) and non-White ethnicity (minority) categories. Women made up 49.8% of the sample and men 49.9% (one did not declare and one selected Other). Mean age was 32.3 years, SD = 12.1. Full-time employed made up 45.6% of the sample, 11.7% was in part-time employment, 6.7% was unemployed and looking for work, 3.5% was unemployed and not looking for work, 3.1% were retired, and 22% were students. The self-employed formed 6.2% of the sample. 70.4% reported that they were not currently seeking a job and 26.6% indicated that they were, with 3% withholding responses.

The Swedish sample comprised of 183 after screening. Due to relatively low numbers in the available pool of Swedish participants meeting our criteria on Prolific.co, this sample was smaller and we were not able to select for equal representation of sex or ethnicity. Inclusion criteria for the sample was Swedish nationality, Swedish as first language, resident in Sweden, and, as the survey was in English, fluent in English language. Women made up 25.7% of this sample, and men 71.6% (one did not declare and three selected “Other”). Mean age was 28.3 years, SD = 8.3. A total of 37.2% of the sample was full-time employed, 13.1% was in part-time employment, 11.5% was unemployed and looking for work, 3.3% was unemployed and not looking for work, 1.1% were retired, and 29.5% were students. Those in self-employment formed 4.4% of the sample. 68.9% reported that they were not currently seeking a job and 28.4% indicated that they were, with 2.7% withholding responses.

Materials, design and procedure

We employed a cross-sectional self-report survey design. The survey measures used in the study were presented in the order of description below using Qualtrics software. Attitudes towards EM wereas measured through 3 aspects, attraction to employers that use EM, credibility of employer use of EM, perceptions of EM in general. Note that items measured attitudes towards the principle of use of EM by employers and respondents were not asked to bring to mind a past or present employer. All scale items had a 100-point slider scale response format (0 = strongly disagree, 100 = strongly agree) unless stated.

Potential participants were initially informed that the study was on perceptions of equality and diversity data monitoring by employers, and that the results will be used to advise real organisations and inform research and policy on the collection of equality data.

Text description of EM

As it is unclear if the public know about or can put a name to EM, especially in Sweden, participants first read background material explaining what EM is and how it is used by organisations with examples from public sector organisations in the UK, differentiating EM from positive discrimination. They were asked to read through all the information thoroughly for understanding. This was followed directly by the two comprehension/attention check questions to check if they had read and understood the text.

Comprehension/attention check

An open-ended question asked respondents to briefly describe EM in their own words. A second question asked if, according to the text, EM is the same thing as positive discrimination (the answer is clearly stated as “no” in the text) with a yes/no response choice. Participants were deemed to have passed the first question if their answers demonstrated an understanding of EM as described, that it involved information or data collection on background characteristics (such as ethnic groups, women or minorities) and/or was a method to address equality and discrimination, and this had to be in their own words rather than copied and pasted from the text. Due to such a variety of verbal responses, exclusion criteria were quite lenient and only those responses that were irrelevant or unclear were excluded on this basis. In total 53 (5.7%) were excluded from the analyses for having failed one or both of these two question checks.

Past experience of EM

One item asked how many times they had completed an EM form when applying for a job, and they chose from one of six ordinal responses, Don’t know, Never, Sometimes, About half the time, Most of the time, Always.

Attraction to employers that use EM

Nine organisational attraction items followed, adapted from the scale developed by Highhouse et al. (2003). These represented three dimensions of attraction: general attraction, job pursuit intentions, and prestige, and Highhouse et al. confirmed that these three sub-dimensions measure a global factor of organisational attraction. Cronbach’s alpha reliability for the nine items was 0.91. The original scale items are worded for rating a specific organisation, and this wording was modified slightly in order to apply to the perceptions of employers in general that use EM. For example, the original item “This employer probably has a reputation as being an excellent employer” was rewritten as “Employers that collect equality data probably have a reputation as being excellent employers.”

Credibility of employer use of EM

Two items were devised to measure the credibility of employers using EM: “Employers that collect equality data seem genuinely committed to equality and diversity” and “I think employers that collect equality data are only interested in equality and diversity for the sake of appearances.” When the second question was reversed scored, the reliability of the two items was low, with Cronbach’s alpha of 0.52, and they each appear to measure separate aspects of credibility. Hence, they were analysed separately.

Perceptions of EM generally

Perceptions of EM itself were measured in six items, with two items specifically on personal data issues (5 and 6):

1. I think that monitoring anonymous statistics in job applications is a good thing

2. I think more organisations should collect data to monitor if they are achieving equality and diversity in their recruitment

3. I think that anonymous statistics can be used to promote fairness in recruitment

4. I think that anonymous statistics can be used to promote inclusion in recruitment

5. I am comfortable providing anonymous data to recruiting organisations

6. I worry that anonymous data from job applicants could be misused by recruiters

Cronbach’s alpha was 0.87 for the 6 items (item 6 reverse scored).

Pro equality and diversity attitudes

As individual attitudes and values have been found to influence positivity to HR diversity infinitives, respondents rated in two items the importance to them of equality and diversity in an employer: “How important is it for you that your employer actively strives towards equality in the workplace?” and “How important is it for you that your employer actively strives towards diversity in the workplace?” Responses were on a scale of 0 = Not at all important, 100 = extremely important. Cronbach’s alpha for the two items was 0.78.

Perceived history of employment discrimination

The next section asked participants to rate the frequency of their past experiences of employment discrimination for each of the seven categories protected in law (age, disability, ethnicity, sex, gender identity, religion, sexual orientation) on slider scales of 0 = Never/Not at all, 100 = All the time.

Demographic and background variables

Several questions asked about demographic and biographical variables including ethnicity, age, and sex, education, current employment.

A final open-ended question invited respondents to add any comments.

Statistical analysis

Ethnic group was categorised as non-White British respondents versus White British based on Prolific pre-screening data in the UK sample. Mean scores were calculated for organisational attraction (9 items), perceptions of EM in general (6 items with item 6 reverse-scored), and pro equality and diversity attitudes was the sum of the scale two items. Credibility of employer use of EM was analysed as two items separately. Data were analysed with t-tests on mean differences, correlations and regressions, and the country comparison between the UK and Sweden with ANCOVAs controlling for sex of respondent. Significance levels were set at p < 0.05, and all alphas reported are two-tailed. Analyses were conducted using SPSS version 26 and regressions utilised the Hayes Process macro.

Results

UK sample

There was a moderate to high correlation between organisational attraction to employers using EM and perceptions of EM generally, r = 0.61, p = <0.001, N = 741. Pro equality and diversity (Pro E&D) attitudes were positively related to attraction and to general perceptions of EM, both r = 0.45, p = <0.001, N = 741, and to credibility of use of EM, r = 0.30, p = <0.001, N = 741, and negatively related to scepticism about employer use r = −0.14, p = <0.001, N = 741.

Perceptions of EM: UK ethnic group effects

Minority and majority ethnic groups differed little on the general perceptions of EM, but ethnic minorities were more likely than majority respondents to rate employer use of EM as cosmetic (… for the sake of appearances), t (739) = 3.52, p = <0.001 (see Table 1). Holding Pro E&D attitudes was higher in the minority ethnic group, t (739) = 3.73, p = <0.001. As the Pro E&D attitude measure was very negatively skewed, however, with over 50% of the sample scoring the maximum on the two question items (61.6% of minority group and 49.2% of majority group scoring both items at max), it was unsuitable for moderation analysis.

Organisational attraction, history of ethnicity-based employment discrimination and UK ethnic group

A linear multiple regression was conducted with ethnic group, history of ethnic-based discrimination and the interaction as predictors. Organisational attraction (mean of 9 items) was the outcome variable. Ethnic group correlated negatively with history of employment discrimination on ethnic grounds r = −0.51, p < 0.001, N = 739, showing that minorities were more likely to report higher frequency of past discrimination.

The overall model was not significant, and ethnic group did not independently predict attraction to employers that use EM, and neither did perceived history of discrimination on the grounds of ethnicity, but the interaction was significant. Simple slopes analysis shows slopes for ethnic group predicting attraction at each level of the moderator. There is a seemingly flat slope for minority ethnic groups for all levels of history of discrimination, but for majority ethnic group respondents, higher levels of perceived history of ethnic discrimination predicts lower attraction scores (left half of Figure 1), though this is based on relatively few cases. In other words, White British respondents with higher perceived history of discrimination found EM use by employers less attractive.

Perceptions of EM in the UK sample: sex differences

Women gave higher ratings than men overall. T-tests show significant differences on all measures at p < 0.01 with the exception of credibility of employers using EM item 2 (… for the sake of appearances) (p = 0.73). Descriptive statistics are shown in Table 4.

Attraction, history of sex-based employment discrimination and sex, UK sample

A linear multiple regression was performed with participant sex, history of sex-based discrimination and the interaction as predictors and organisational attraction as outcome (regression statistics displayed in Table 2). Sex of respondent correlated negatively with attraction, r = −0.12, p = <0.001, and with history of sex-based employment discrimination, r = −0.28, p = < 0.001, women had higher attraction scores, and reported higher frequency of past sex-based discrimination.

Sex of respondent was a significant independent predictor of attraction to employers who use EM, with women having higher attraction ratings than men. Perceived history of sex-based discrimination did not predict attraction scores. Simple slopes analysis for the significant interaction does suggest that higher perceived history of sex-based discrimination moderates attraction scores in males, whereas perceived history of sex-based discrimination in women had little effect on attraction, with a seemingly flat slope (right half of Figure 1). In other words, male respondents with higher perceived history of discrimination found EM use by employers less attractive. To check if this result was driven by the same participants as the ethnicity result, the regression was rerun with the interaction between ethnic group and sex of respondent included. The interaction between sex of respondent and ethnic group was indeed significant, B = −6.24, SE = 2.48, β = −0.16, t = 2.51, p = 0.01. Examination of the split correlations indicated that the negative relationship between history of sex-based discrimination and attraction was significant only within the White male subgroup (r = −0.20, p = 0.01, N = 192) and not the minority male subgroup (r = 0.11, p = 0.16, N = 176).

Comparison of UK and Sweden samples

The majority of UK respondents had at least sometimes completed EM forms for job applications in the past (76.5%), and over 53% responded with either Always or Most of the time. Only 16.7% had never completed an EM form for a job application. In contrast, only 23% of the Sweden sample had at least sometimes completed EM forms for job applications, while the majority 67.8% indicated they had never done so [2].

Descriptive statistics on attitudinal measures for the two country samples are presented in Table 3 and all group means differed significantly. UK respondents gave higher ratings on everything except for the second credibility question, where the Sweden sample showed slightly higher ratings reflecting a more sceptical view of employer use of EM.

Due to an overrepresentation of males in the Swedish sample (over 70% male), analysis of covariance (ANCOVA) tests was conducted with sex of respondent as the covariate to control for any effect, and with perceptions of the five aspects as dependent variables (Table 4). Sex was a significant covariate in all cases while the country sample was significant when sex was controlled for except in the case of EM credibility item 2 where there was no effect of either. The UK sample had more favourable perceptions of EM and rated the importance of equality and diversity in an employer higher than a Swedish sample.

Two of the 6 items on perceptions of EM in general measure how comfortable/concerned respondents are in providing personal data. These are tested separately below as it is especially informative to see country differences. The UK sample had higher scores on the item “I am comfortable providing anonymous data to recruiting organizations”, M = 81.78, SD = 21.91 compared to the Sweden sample, M = 69.89, SD = 30.88, and this was significant with Welch’s t-test for unequal samples, t (229.19) = 4.91, p < 0.001, Cohen’s d = 0.496. For the second item about personal data “I worry that anonymous data from job applicants could be misused by recruiters” the difference was also significant but smaller with the Sweden sample giving higher ratings, M = 53.99, SD = 30.49, than the UK sample, M = 46.10, SD = 29.57, Welch’s t (272.78) = 3.15, p < 0.001, Cohen’s d = 0.265. The Sweden sample was overall less comfortable and more concerned with providing personal data for EM.

Discussion

There are potential benefits for organisations and society in collecting and monitoring data on equality and diversity, and it is increasingly a requirement of governments for organisations to provide at least some data. Surveys within the EU over the past two decades have shown evidence suggesting public support for data collection in the field of employment, but in many countries there is patchy EM – even if data collection exists it is not fully utilised. Our study is to our knowledge the first to measure detailed perceptions of EM. The results hint at cross-sectional differences depending upon country setting and individual factors. While the cross-sectional design does not permit us to make causal inferences, this topic is exploratory and establishing differences among different populations and relationships between variables is a useful first step in determining where to look for causality.

Employer use of EM and the perceptions of EM in general were positive, and this was the case for all groups and across country samples. In relation to RQ1, “Do UK minority and majority ethnic groups differ in their attitudes towards EM?”, the analysis of UK majority and minority ethnic groups revealed no differences overall in attitudes towards EM itself. However, minority group respondents tended to be more sceptical in their views of employer use of it, and more likely to agree with the statement that employers use EM for the sake of appearances. The fact that EM is prominent in the UK, and employers are mandated to engage in the process to some extent, will inevitably mean that the motivation may be seen as mere compliance with legal obligations or for purely cosmetic reasons. If EM is very commonplace this will not affect organisational attraction as it is seen as standard HR procedure. However, its absence might be noticed in minority groups and be seen as an indicator of a non-inclusive workplace. A future experimental study which manipulates the presence versus absence of EM should test this directly.

Addressing the question of “Does history of discrimination interact with ethnicity to effect attraction to organizations using equality monitoring?” (RQ3), the attitudes of minority non-White respondents towards EM were not tied to perceived past experiences of ethnic discrimination in employment. The interesting result here was that majority White British respondents who reported a higher level of past ethnic discrimination were less positive about employer use of EM. This pattern was repeated in the analysis of sex. Men who perceived a higher level of past sex-based discrimination were also less positive towards employers using EM, though this result was driven by a subgroup of White males. It is likely that these results reflect fears in majority respondents about losing out to token diversity hires and represent a threat to self-interest. Several explanations have been put forward in the literature to account for why majority group members are often less positive about diversity initiatives. Plaut et al. (2011) propose that some whites feel excluded by multiculturalism with its emphasis on minority cultures. This threatens the need for inclusion and is a source of anxiety, hence the appeal of colour-blind initiatives over identity-conscious ones often found in the literature. Our background text on EM was identity-conscious, listing examples of underrepresented groups, and this may have heightened a sense of threat or exclusion for majority respondents who hold beliefs about previous discrimination. This could also indicate concerns about fairness. Kravitz (2008) reports that perceived fairness of affirmative action diversity initiatives is a better predictor of how initiatives are perceived than self-interest, and is sometimes a genuine motivation, though it can also be used as a rationalisation for prejudice. Nonetheless, we cannot rule out that prejudice or intolerance may be the underlying factor that explains our findings. Majority group member perceptions of past discrimination, which should be statistically rare, could well be an indicator of underlying dislike or mistrust of multiculturalism (Ryan et al., 2007), racism, intolerant attitudes to women, or authoritarian tendencies (Strauss et al., 2003). However, our findings do support ideas that majority group respondents as a whole do not react negatively to EM and suggest that other factors mediate this. Further research that determines causal inference is needed to unpack this and provide insight into which of above explanations play a role.

Regarding RQ2 “Are there sex differences in attitudes towards EM in the UK sample?”, women were more positive about EM and about employers that use it, and had higher pro equality and diversity attitudes. This supports previous research that women are more positive and engaged with workplace diversity initiatives than men overall (Soni et al., 2010).

Finally, when addressing the question of “How do UK and Swedish samples differ in their attitudes towards EM?” (RQ4), as compared to Sweden, the UK sample had more favourable attitudes towards EM and employers that use it, and held more pro equality and diversity attitudes, even when sex, which was skewed in the Swedish sample, was controlled for. The Sweden sample was more concerned about, and less comfortable with providing personal data for EM than the UK sample. However, the figures suggest that Swedes were still on average comfortable providing such data (the mean was around 70 on a 100-point scale), although they did indicate moderate levels of worry about the potential misuse of data by recruiters (mean of 56). There may well be less trust in anonymised processing of equality data in the Swedish employment sector, and standardised job application forms and anonymised online recruitment systems are less prevalent than in the UK, and applications by CV and cover letter are more common. As noted in the introduction, it is unusual for such personal data to be collected in Sweden and data on ethnicity may be a particularly sensitive issue in the current political climate.

The text on EM presented to respondents at the beginning explained what EM is and presented it as a tool to aid in inclusion, and therefore cast it in a favourable light. It is possible that participant responses were influenced by demand characteristics, and social desirability has been known to inflate ratings in studies on equality and diversity. Thus, it may be difficult to translate the strength of those ratings to real life attitudes. However, the group differences revealed in our results should be unaffected by these issues.

Since we have established these differences between the countries, the extent to which UK and Swedish respondents might more or less favour identity-conscious or identity-blind diversity initiatives would be pertinent to examine in future research. Sweden has a shorter history of large-scale immigration than the UK, and questions of diversity and equality in the country are more often focused on sex and gender than ethnicity. The concept of equality and sameness is arguably deeply enshrined in Swedish culture through “Jante Law”, a set of values believed to influence Scandinavian countries, and which emphasises how no person is different to or better than another. It may be that identity-conscious diversity initiatives like EM, with their focus on inclusion of disadvantaged and minority groups, are culturally less attractive in countries influenced by these values. There is little research addressing cross-cultural differences in preferences for colour-blind approaches over identity-conscious ones, but this is an interesting avenue to explore in future work. As the Sweden sample was biased towards being slightly younger and was over 70% male, future studies should seek out a more representative and diverse sample. However, as an initial exploratory study, these findings provide impetus to further research.

Implications for research and practice

Though lower than the UK sample, the average attitudinal ratings of the Sweden sample were on the positive side of the scale, and despite moderate concern about the misuse of personal data, it does suggest that the principle of EM, at least when it is explained with examples, is favourably received. There is a high level of trust in government among the public in Sweden (OECD, 2020). This bodes well for the expansion of EM if the government chooses to embrace it. The Public Sector Equality Duty has likely reinforced the legitimacy of and trust in EM in the UK. It will be important for governments to lead the way in both collecting and using equality and diversity data, providing guidance to organisations to engage with monitoring as part of diversity management, if this practice is to become more widespread.

Diversity and inclusion branding is seen as conferring a competitive edge, influencing attraction and impacting upon hiring and talent management through signalling person-environment fit (Byrd, 2018; Jonsen et al., 2021; Ng and Burke, 2005; Younis and Hammad, 2021). Strategic marketing of EM could enhance employer and corporate brand to demonstrate real and active commitment, walking the talk, that backs up value and culture statements. If data and aspects of the monitoring process are made available, on corporate websites and in recruitment materials, it creates transparency, providing an authentic picture of organisational climate. Authentic messages about organisational culture, even if it is less diverse or inclusive than it strives to be, may be better perceived than inauthentic statements contradicting reality and falsely claiming diversity and inclusivity (Cording et al., 2014; Marcinko, 2020). For EM to be genuine, it must go beyond compliance or mere image management, and actively inform and evaluate diversity management policies. The limited evidence suggests the equality data are currently not being fully utilised in the fight to increase inclusion, even where such data collection is widespread (Hannigan et al., 2019). We argue that, used this way, it can augment other efforts towards inclusive workplace climate and culture (Le et al., 2021), which itself has considerable benefits for the organisation (Chung et al., 2021). EM could therefore help to erase the gap between values and organisational reality, fostering an evidence-based approach to diversity management.

Compliance with data handling is mandated in the European data protection law (GDPR) and completion of EM forms is voluntary and anonymous. Forms are separated from job applications and not allowed to bias the managers responsible for selection, with this anonymity communicated clearly to reassure respondents. This may provide challenges for small companies or those that presently use email or less formal application systems, which are more common in Sweden than the UK. Care must be taken to communicate why data are collected and how the resulting data are used to allay the fears of some majority group members who feel excluded by language that focusses on disadvantaged groups, and, in particular, to avoid confusion with strong affirmative action or positive discrimination measures like quotas. The latter are, in practice, rarely used as they breach most countries’ anti-discrimination laws. Yet the association with diversity initiatives like EM may well pervade in some members of the public and cause some majority groups to perceive unfair treatment where it does not exist.

Conclusions

Notwithstanding the effects on certain majority group members, EM and its use by employers was viewed positively even in a sample from a country where EM has currently low prevalence. Given the potential usefulness and importance of equality data collection in recruitment, both for organisations, government and society, the results support a greater use in diversity management. It would be beneficial to experimentally manipulate EM use by employers and examine the effects directly on outcomes such as organisational attraction, job-pursuit intentions, and “employer of choice” decisions (Tanwar and Kumar, 2019). While some majority group members may find EM diversity initiatives less appealing, it remains to be seen if they have effects on real job choice decisions, or if their absence might impact attraction and job pursuit in minority groups. Person-environment fit and signalling approaches might suggest it does. On the basis of our findings, these questions merit further study.

Figures

Figure 1

Scatterplots of organizational attraction by perceived personal history of employment discrimination with best fit lines, for both ethnicity (left) and sex (right) of respondent

Table 1

Descriptive statistics by ethnic group and sex of respondent, UK sample

	Ethnic group
	Minority (N = 362)			Majority (N = 379)
	M	SD	95% CL	M	SD	95% CL
Organizational attraction	69.68	17.15	67.91, 71.45	70.72	16.96	69.01, 72.43
Attitudes to EM generally	73.37	16.37	71.68, 75.06	74.71	18.66	72.82, 76.59
Pro E&D in an employer	172.46	35.69	168.77, 176.15	161.40	44.40	156.92, 165.89
EM credibility 1: genuinely committed to equality and diversity	61.53	24.64	58.98, 64.08	63.11	25.82	60.50, 65.72
EM credibility 2: only for the sake of appearances	58.64	23.02	56.26, 61.01	52.44	24.90	49.92, 54.95

	Sex of respondent
	Female (N = 370)			Male (N = 371)
	M	SD	95% CL	M	SD	95% CL
Organizational attraction	72.35	16.81	70.63, 74.07	68.08	17.04	66.34, 69.82
Attitudes to EM generally	76.80	15.15	75.25, 78.35	71.32	19.34	69.35, 73.30
Pro E&D in an employer	180.02	29.87	176.97, 183.07	153.63	45.60	148.97, 158.28
EM credibility 1: genuinely committed to equality and diversity	65.11	24.49	62.60, 67.61	59.58	25.71	56.95, 62.20
EM credibility 2: only for the sake of appearances	55.77	24.96	53.22, 58.32	55.16	23.41	52.77, 57.55

Table 2

Regression statistics for the effects of ethnicity, sex of respondent and history of employment discrimination on attraction to organisations that use EM

	Ethnicity and history of ethnicity-based discrimination
	B (SE)	β	t
Constant	69.45 (0.99)	Na	70.48***
Ethnic group	−0.60 (1.65)	−0.02	0.36
History of ethnic discrimination	0.18 (0.03)	0.03	0.69
Interaction	−0.15 (0.07)	−0.10	2.00*
Total R2	0.006
F (3, 735)	1.55

	Sex and history of sex-based discrimination
	B (SE)	β	t
Constant	72.24 (0.90)	Na	80.20***
Sex	−4.82 (1.29)	−0.14	3.73***
History of sex discrimination	0.15 (0.03)	0.02	0.50
Interaction	−1.11 (0.05)	−0.12	2.50**
Total R2	0.026
F (3, 734)	6.41***

Note(s): Ethnic group coded as minority = 0, majority = 1. Sex coded as female = 0, male = 1. *p < 0.05 (2-tailed). ***p < 0.001 (2-tailed)

Table 3

Descriptive statistics on attitudes to equality monitoring by country sample, UK and Sweden

	Country sample
	UK (N = 741)			Sweden (N = 183)
	M	SD	95% CI	M	SD	95% CI
Organizational attraction	70.20	17.05	68.98, 71.44	59.08	18.44	56.39, 61.77
Attitudes to EM generally composite score	74.06	17.58	72.79, 75.32	61.93	23.31	58.53, 65.33
Pro E&D in an employer	166.80	40.73	163.87, 169.74	135.25	52.27	127.62, 142.87
EM credibility 1: genuinely committed to equality and diversity	62.34	25.24	60.52, 64.16	53.16	25.30	49.47, 56.85
EM credibility 2: only for the sake of appearances	55.46	24.18	53.72, 57.21	58.83	23.70	55.37, 62.28

Table 4

Analyses of covariance test statistics, country comparison controlling for sex of respondent

	Covariate of respondent sex			Country effect (UK vs Sweden)
	F (1,918)	p	ηp2	F (1,918)	p	ηp2
Attraction composite score	19.39	<0.001	0.021	47.32	<0.001	0.049
Attitudes to EM generally composite score	25.41	<0.001	0.027	45.62	<0.001	0.047
Pro E&D in an employer	107.34	<0.001	0.011	54.85	<0.001	0.056
EM credibility 1: genuinely committed to equality and diversity	11.39	0.001	0.012	13.39	<0.001	0.014
EM credibility 2: only for the sake of appearances	0.038	0.846	<0.001	3.08	0.079	0.003

Note(s): Country was coded 0 = UK, 1 = Sweden

Notes

1. A small number were excluded for suspiciously fast responses (under 3 min), discrepancies between Prolific pre-screen data and answers to our demographic questions, or failures of the attention check (see below).

2. Note: an analysis of group differences based on previous experience of EM across the sample were not significant and therefore left out of the article for brevity.

Declaration of competing interest: The authors declare that they have no known competing interests.