In this page:

• Why was the practice needed?
• How was it implemented?
• Outputs
• Key success factors
• Technical information
• Contacts

Why was the practice needed?

Discrimination testing is considered to be the best methodological design to objectively detect actual discrimination. Together with data from the survey on experiences of discrimination in the Netherlands and on discrimination complaints (see related practice in this compendium), this type of controlled field experiment in real-life settings can provide a comprehensive perspective of discrimination.

The Dutch Ministry of Social Affairs has conducted several studies that use job application-based discrimination testing. In 2008, the Netherlands Institute for Social Research conducted a large-scale test whereby it sent out fictitious job applications for low, middle, and high level positions under fictitious ‘native’ Dutch and foreign sounding applicant names in order to measure the extent of ethnic discrimination in the jobs market.

In 2010, they conducted similar testing, but instead targeted job agencies and used actors to represent applicants.

In its 2014 Action Plan on Discrimination in the Job Market, the Dutch Government set out to combat discrimination and called for further research into this type of discrimination and the effectiveness of anti-discrimination measures.

Responding to this need, in 2015, the Ministry of Social Affairs and Employment commissioned Panteia, a research firm, to conduct online discrimination testing. The results are used by the government to raise awareness and to inform the policy cycle in consultation with NGOs and employers’ associations.

How was it implemented?

The discrimination testing consists of posting fictitious resumes and CVs on online job portals to assess whether employers or third-party recruiters reach out to potential candidates based on their age, ethnic background or gender.

Researchers uploaded a total of 1,742 fictitious job profiles.

The test involved a control group of fictitious applicants, consisting of ‘native’ Dutch men aged between 23-35. In addition, the research created seven experiment groups each consisting of applicants who are of Turkish, Moroccan, Antillean, Surinam or Polish origin, women, and ‘older’ applicants (aged 40-45).

The respective job-profiles of the control and experiment groups were kept identical, save for the discrimination ground. Researchers monitored the amount of clicks each job profile received in addition to the amount of responses from employers or third-party recruiters (emails, voicemails and missed calls) received by the fictitious applicants.

Researchers conducted tests in and around eight major cities in the Netherlands, and focused on six occupations, ranging from low to higher skilled job profiles. The employer category consisted of ‘actual’ employers and third-party recruiters.

In contrast to previous discrimination testing, the study conducted by Panteia covered various grounds of discrimination.

Following a design developed by Blommaert, Coenders and Van Tubergen (published in the academic journal ‘Social Issues’ in 2014), it was possible to monitor employers’ or recruiters’ reaction in two subsequent stages. The first was whether employers or recruiters – after seeing only limited information, such as the name of the job applicant – inspected the resume of the applicant. The second was whether employers or recruiters contacted the job applicant.

Outputs

• A report (2015) setting out the research design and results. The results have fed into the policy cycle, e.g. with regard to governmental Action Plans to combat discrimination.

Key success factors

• Discrimination testing by means of field experimental studies have become popular as they provide convincing evidence of actual discrimination. They are applied in real-life settings (e.g. the labour or housing market) and indicate actual discriminatory practices. The experimental design of such studies (with different experiment groups) allows causal conclusions to be drawn. Generally, field experimental studies are considered the best method to detect actual discrimination.
• There are important methodological issues to bear in mind when conducting discrimination testing. For one, the design of the application materials should be realistic and valid. For example, in a study of labour market discrimination, the fictitious CVs and motivation letters should be realistic and tailored to the type of vacancies that are targeted. Save for the possible discrimination ground (e.g. the applicant’s ethnic background), the CVs and motivation letters should be identical or equivalent.
• Another issue regards ethical concerns. Acquiring informed consent is not possible in this type of research, as employers should be unaware that the job applicant is fictitious for the research to be successful. Given the societal relevance and the inability of detect discrimination well by other methods, ethics committees will generally approve this type of design.
• However, careful attention should be paid to minimise the burden for employers. At a minimum, when a fictitious job applicant receives a positive reaction from an employer, the researchers should respond to the employer as soon as possible to inform that the applicant is no longer available for the position.

Technical information

• Data sources covered: Discrimination testing.
• Areas of life covered: Employment.
• Target audience: Public authorities, labour inspection, researchers, and the general public
• Duration: One research conducted in 2015, another on the way in 2018/2019
• Geographical scope: National
• Leading institution: Ministry of Social Affairs and Employment
• Other organisations involved/consulted:The Netherlands Institute for Social Research and Panteia (research company)
• Financial & human resources: No information provided.

Contacts

The Netherlands Institute for Social Research / Sociaal en Cultureel Planbureau

Email: info (at) scp (dot) nl

Panteia

Email:info (at) panteia (dot) nl