The difficulties in comparing profile variability (a measure of test scatter) are briefly discussed and the limitations of current techniques pointed out. Test scatter is defined as individual variation in test scores between or within various psychological and educational tests. Currently, no statistical technique for the comparison of profile variability across groups is available. Profile variability is explained in detail and a statistical technique is developed to compare profile variability across independent groups, i.e., the variance of each subscale or subtest within an instrument. Although development of the statistic is accomplished with applications to the Wechsler Intelligence Scale for Children (Revised) in mind, it is generalizable to any set of test scores derived for an individual. Theoretical and practical applications of the new technique to diagnosis, and description of abnormal populations are discussed. (Author/RL)