The assessment of the achievement of students and the quality of schools has drawn increasing attention from educational researchers, policy makers, and practitioners. Various test-based accountability and feedback systems involving the use of value-added techniques have been developed for evaluating the effectiveness of individual teaching professionals and schools. A variety of models have been employed for calculating value-added measures, including the use of linear regression models which link students' present and prior achievements. One of the limitations associated with the use of the conventional linear regression methods in value-added calculation is that the value-added measures are likely to be overestimated for students with higher prior achievements while underestimated for students with low prior achievements. This study explores an alternative approach, the principal axis approach, to calculating value-added measures which can eliminate some of the limitations associated with the conventional linear regression approaches.