Several influential studies show that living in more ethnically diverse settings erodes generalized social trust among the native-born population. Yet, critics argue that this is a biased portrayal of the consequences of ethnic diversity for trust because co-habitation only entails superficial, stereotype-confirming exposure to other ethnic groups rather than meaningful contact with them. Conversely, interethnic contact can in fact build up, rather than erode, generalized social trust, according to this perspective. However, putting the “contact argument” to a convincing empirical test has been complicated by several methodological challenges. In this paper, we provide a rigorous test of this argument by studying a contact-prone context—the workplace—using registrylinked survey data from Denmark. Despite setting the scene for the contact argument, our empirical analyses consistently show a negative effect of ethnic diversity in the workplace on social trust. This result holds in a series of analyses, including in panel data models. Our results thus provide little support for the contact argument, and instead vindicate theories highlighting the negative consequences of ethnic diversity for trust.