A test for racial discrimination | FAQs

FAQ: Study

How did you decide which names to use?

An initial list of names was drawn up based on those found in survey data we held and that were common within particular ethnic groups and relatively exclusive to them. A quantitative assessment of their association with ethnic groups was carried out using an on-line omnibus panel. Each name was tested with a minimum of 650 adults of working age in Great Britain.

What were the occupations that you applied for?

We sent jobs applications to nine different occupations. These occupations were: IT Support, IT Technician, Accountant, Accounts Clerk, Human Resources Manager, Teaching Assistant, Care Assistant, Sales Assistant and Office Assistant. Discrimination against ethnic minority groups was evident across all of these occupations

How did you choose which jobs to apply for?

advertised on job-search websites and in the local newspapers of the cities that were taking part in the study. Eligible vacancies were defined using job descriptions that included possible job titles, roles and level of seniority.

How were the applications developed?

A set of three applications were developed in response to each job advert. The applications were matched closely in terms of their education, skills and work history. This meant that the three applicants would appear comparably qualified for the advertised position. All applications were created to reflect plausible candidates who were sufficiently different as to not arouse suspicion about the study. In cases where specific skills (such as IT programming skills) were mention in the job advertisment, all three applications would be adjusted to increase the likelihood of a positive response from the employer.

Was the level of discrimination statistically significant?

Yes. Of those instances where at least one of our applications met the required standard, the net level of discrimination in favour of white applicants was 29 per cent. The 95 per cent confidence interval around this figure was 19 to 39 per cent. You can find more information about these statistics in the full report.

How can you be sure that the applications were equivalent?

We put a great deal of effort into making them as similar as we could. However, the random allocation of the name at the end of the development process means that over the study as a whole, even if there were differences between applications, the effect of this would be about equal between the ethnic groups.