The project provides a new state-of-the-art scalable methodology to estimate various preferences about discrimination on large scales for different domains, e.g. for discrimination based on gender, ethnicity, and LGBTQ+ status. Public and private institutions can use this methodology to estimate discrimination preferences. In terms of impact, this will allow a better understanding of discrimination, e.g. understand the extent of discrimination, which type is more prevalent, and who discriminates, and serve as a basis for evidence-based policy to alleviate discrimination. In terms of methodology, preliminary results indicate that the methodology has good technical properties, such as consistency of answers in line with what is expected from the theoretical framework. Regarding the UK estimation, preliminary results indicate that taste-based preferences exist among a minority and that most individuals engage in statistical discrimination. Most have a moral-based preference against taste discrimination, and, unlike in classical models, most of the statistical discrimination in the experiments is explained by a moral-based preference to reward the most-productive-on-average group. We also produce various other results, such as regarding who tends to discriminate and how discrimination preferences map into policy and political support.