In recent years, various studies have been carried out to investigate ethnic discrimination. Compared to other countries, however, the findings for Germany - especially in the field of experimental research - are still sparse. My work aims to close this research gap by investigating the extent to which discrimination can be attributed to preference-based or statistical discrimination. In a factorial survey design, I use written CVs with applicant pictures of different ethnic origins (German / Asian / African) as well as corresponding application videos created by deepfakes. The latter enable the targeted manipulation of the ethnic appearance of the applicants while using an identical audio track. Based on an online survey of people with personnel responsibilities (n=1,316), the results show no evidence of preference-based discrimination. Furthermore, more information through application videos even leads to a significantly higher probability of non-German applicants being invited to an interview. Even if employers suspect that certain skills, such as German language skills, are lacking, these are shown in the video and can be considered. Video applications could therefore be a suitable means of convincing potential employers of one's own abilities and not being assessed on the basis of assumed characteristics.