The meaning of standardized achievement testing to 50 third-grade teachers was investigated through a construct validation procedure using questionnaires and structured open-ended interviews. Subjects were from 18 schools performing distinctly above or below prediction on the Delaware Educational Assessment Program Longitudinal Study. Multimethod-multitrait procedures show the construct, evaluation acceptance, as valid, consistent, and distinct from trait anxiety. The meaning underlying the construct has dual evaluative dimensions, unanimously salient potency dimensions, and a deficient knowledge base. Implications are proposed for teacher education, state assessment models, administrative adjustments, and research on causal influences of teacher attitudes on testing. (Author/CTM)