More importantly in the analysis sample only one group, with 13 observation, fall below the recommended level of 20 cases per group. Although the group size would exceed 20 if the entire sample were used in the analysis phase, the need for validation dictated the creation of the holdout sample. The three group are of relatively equal sizes , thus avoiding any need to equlize the group sizes. The analysis proceeds with attention paid to the classification and interoretation of this small group of 13 observation