The validity of grades given by secondary school teachers was studied. Two teacher-originated measures of performance and one independent standardized measure of achievement were examined. The teacher-originated grades were final examination grades (FINALs) and end of the semester marks (MARKs) for over 40 courses at all secondary levels. The standardized tests were from the "Survey Tests of Essential Elements/Learner Standards" (STEELS) covering the instructional content of the more than 40 courses studied, spanning the areas of language arts, mathematics, social studies, and science. Over the 44 tests examined, correlations among the three measures--MARKS, FINALS, and STEELS--often varied markedly among each other both within and among tests. No strongly consistent patterns were observed; only weak trends and tendencies were found. Within-teacher and between-school validity coefficients tended to be greatest in freshman level courses with large enrollments, which were heterogeneous in competency. There were no clear differences in marking and grading across disciplines. The generally low validity coefficients for teacher final examinations and end of semester grades imply that these measures reflect many characteristics of students. An example of the marking and grading systems typical of those used by teachers in the Dallas school system (Texas) illustrates the difficulties of marking systems, particularly when the student receives a zero for incomplete work. Ten tables provide data on student performance. (SLD)