In longitudinal or cross-sectional cephalometric evaluation there are many factors which make analysis difficult. Samples are often limited in number and include both sexes. Registration may be done at different ages, using different equipment and techniques. By linear scaling size differences between sub-groups may be calibrated. Angular variables may still differ, however, suggesting differences in shape. If not, such sub-groups can be pooled, thus increasing the number of cases in each group and allowing more meaningful comparisons to be made with other groups. Calibration of biological size differences and/or technical effects of image enlargement may also be used to create a suitable baseline independant of growth, from which to study treatment effects.