A comparison of Sheldon's and Parnell's methods for quantifying morphological differences

Comparison of Sheldon's somatotype and Parnell's phenotype scores on 102 normal male university students showed significant differences between all three matched pairs of components. Shared variances were 35% for endomorphy and Fat, 44% for mesomorphy and Muscularity, and 74% for ectomorphy and Linearity. Parnell's Muscularity scores averaged nearly one point less than Sheldon's mesomorphy scores. Although both systems exhibit redundancy, neither seems reduceable to two scores. Because both systems show twice as many correlations with a variety of psychological measures as would be expected by chance, it would be useful to include either or both systems routinely in psychological studies.