The persistent gap between the performance of mainstream students and racially and linguistically diverse students--for example, African Americans, Hispanic Americans, and Native Americans--on standardized tests may well signal problems with procedures for the development and use of standardized tests in general, and for their use with culturally diverse populations in particular--problems made more or less invisible by a lack of serious attention to social perspectives on language and cognition within the field of assessment. For decades, cultural psychologists and anthropologists have been developing ever more sophisticated understandings of the critical role that culture and society play in cognitive and language development. Also for decades, linguists and theorists in the field of language and literacy have been developing ever more sophisticated understandings of the important roles played by readers' prior knowledge and experience in making sense of written texts and the essentially social and dialogic nature of language. Yet the assessment field, firmly rooted in an objectivist tradition, has been slow to attend to the implications of these advances, despite compelling evidence that conflicts of language and culture pose threats to assessment validity, particularly (and typically) for linguistic minority students, but for mainstream students as well. In this article, the author believes that the issue of validity is critical because test scores provide the basis for long-term decisions about students concerning placement, selection, certification, and promotion. These long-term decisions can have significant consequences for students and not all of those consequences are good. (Contains 2 notes.)