When selecting employees for higher or specialized posts employers frequently use written tests. However, such tests may contain a hidden bias which results in certain groups not doing as well as the general population of applicants. The question is whether a test has to be proven to be discriminatory before it can be challenged by employees, or prospective employees. .
Now the UK Supreme Court has ruled that an employer can be sued if it can be shown that one group distinguished by race, colour or ethnic origins performed significantly worse on a test used to select people for jobs. There is no prior requirement to establish why the tests produces such results.
A group of 49 employees working for the UK immigration service have been given leave by the Supreme Court to pursue their discrimination case against the Home Office. This is because in the Executive Officer examinations non-white employees have a success rate that is just 40.3% of that for white candidates.
This case illustrates the dangers of blanket testing, although specific testing for an individual technical or linguistic skill may be defensible if it can be shown that the skill is strictly necessary for a particular job. The fact that any group performs less well than the general population in a screening test for all job applicants could, for instance, reflect the higher aspirations of a particular culture or subculture such that people with less capability seek advancement from within that group. This factor would not, however, be a likely possibility if – for instance – a group of junior executives at a particular level within a company took a common test and one group distinguished by their racial origin performed less well. This is because there is no ethnic (or other) difference in intelligence levels. However, there is a known sexual bias in intelligence tests, as women generally score higher on verbal than non-verbal tests than men.