This suggests an
alternative criterion by which to judge changes in student performance - namely, that achievement gains on test items that measure particular skills or understandings may be
meaningful even if the student's overall test
score does not fully generalize to other exams.