This paper contributes to the existing literature on ethnic discrimination of immigrants in hiring by addressing the central question of what employers act on in a job application. The method involved sending qualitatively identical resumes signalling belonging to different ethnic groups to firms advertising for labour. The results show that whether the applicant has a native sounding or a foreign sounding name explains approximately 77 per cent of the total gap in the probability of being invited to an interview between natives and immigrants, while having foreign qualifications only explains the remaining 23 per cent. This in turn, suggests a lower bound for statistical discrimination of approximately 23 per cent of total discrimination. The analysis indicates further that the 77 per cent are most likely driven by a mixture of preference-based and statistical discrimination.