Two questions motivated this study: (a) Does test familiarity influence teachers' judgments of their students' test performance? and (b) Does the disability status of students influence their teachers' judgments? Teachers (n=19) judged item performances for one student with disabilities and one student without disabilities (n pairs=19) from their fourth-grade classrooms. Teachers made judgments using (a) a mathematics test from the research version of the TerraNova CTBS Multiple Assessments edition, which is similar to the large-scale achievement test administered in numerous states, and (b) classroom-based math tests. Judgment accuracy was higher (a) on classroom tests and (b) for students without disabilities. Among less accurate judgments, teachers consistently underestimated the performances of students with disabilities. Students with disabilities performed lower on both types of tests. Student test performance accounted for a significant proportion of the variance in teacher judgment accuracy. Implications of these findings for future research and practice are discussed.