Empirical contributions to the study of discrimination in the labor market.

Summary This thesis is composed of three articles, each of which asks a question about discrimination. The first article examines whether the inhabitants of Seine-Saint-Denis are statistically discriminated against on the basis of supposed violence . where the practice of a combat sport would be a signal of violence. The second article looks at whether Marseilles accents are discriminated against in the labor market. The last article asks whether employers could use the surname and address of job applicants to correlate their accent and, if they infer that they have an accent from the big cities, to discriminate against them on this point. The "testing" method is used to answer these three questions. The first two testings lead to a negative answer to the first two questions. Regarding the third question, even if it cannot be answered, the third test proves that the accent of the large ensembles is discriminated. Moreover, it shows that discrimination on the patronymic can overshadow that on the address. In general, this work shows that employers do not take sport into account in their hiring practices and that when designing a test, one must distinguish between strong and weak signals. He also shows that testing can be used up to the point where employers contact candidates by phone and engage different stereotypes depending on the accent they hear. Finally, he suggests that policymakers should seek to reduce discrimination based on surname first if they want to reduce total discrimination.