Following on from a cross‐sectional analysis of refraction data (Saunders, 1981), the present paper attempts the evaluation of serial recordings of a new sample of refractive corrections using longitudinal methods. This study compares the results of the cross‐sectional and longitudinal methods, investigates whether or not the latter supports the conclusions of the former, examines to what accuracy this is achieved and, where the results differ, seeks a reason for this. It is demonstrated that in spite of such differences the two methods coincide in their conclusi