This article describes the history, evidence warrants, and evolution of the Center for Research on Evaluation, Standards, and Student Testing's (CRESST) model-based assessments. It considers alternative interpretations of scientific or practical models and illustrates how model-based assessment addresses both definitions. The components of the models, that is, research in support of domain-independent elements and research in support of domain-specific explication, are presented and justified in their application to multiple purposes of testing and assessment and validation thereof. Descriptions of research studies are presented as they clarify conceptions of validity for different purposes, fairness, and appropriate use. Summaries of major findings are given in the areas of test design and scoring and the linguistic demands of tests. The emphasis on transfer and generalization as a major validity element is described, in particular, the use of worked examples supported by schema theory to reduce cognitive load. POWERSOURCE[C], a new CRESST project, is described in the light of its assessment strategy and its support of professional development and use. (Contains 5 figures.)