Racial attitudes, beliefs, and motivations lie at the center of manyinfluential theories of prejudice and discrimination. The extent to which such theories can meaningfully explain behavior hinges on accurate measurement of these latent constructs. We evaluated the validity properties of 25 race-related scales in a sample of 910,066 respondents using various tools,includingdynamic fit indices, item response theory, and nomological nets. Despite showing adequate internal reliability, many scales demonstrated poor model fit and had latent score distributions showing clear floor or ceiling effects, results that illustrate deficiencies in thesemeasures' ability to capture their intended latentconstruct. Nomological nets further suggested that the theoretical space of "racial prejudice" is crowded with scales that may not capture meaningfully distinct latent constructs. We provide concrete recommendations forboth scale selection and scalerenovation and outline implications for overlooking measurement issues in the study of prejudice and discrimination.