The Percentage of Mature Height as a Morphometric Index of Somatic Growth: A Formal Scrutiny of Conventional Simple Ratio Scaling Assumptions.

Abstract

To assess conventional assumptions that underpin the percentage of mature height index as the simple ratio of screening height (numerator) divided by actual or predicted adult height (denominator). We examined cross-sectional data from 99 academy youth soccer players (chronological age range, 11.5 to 17.7y) skeletally immature at the screening time and with adult height measurements available at follow-up. The y-intercept value of -60cm (95% confidence interval, -115 to -6cm) from linear regression between screening height and adult height indicated the failure to meet the zero y-intercept assumption. The correlation coefficient between present height and adult height of .64 (95% confidence interval, .50 to .74) was not equal to the ratio of coefficient of variations between these variables (CVx/CVy = 0.46) suggesting Tanner's special circumstance was violated. The non-zero correlation between the ratio and the denominator of .21 (95% confidence interval, .01 to .39) indicated that the percentage of mature height was biased low for players with generally shorter adult height, and vice versa. For the first time, we have demonstrated that the percentage of mature height is an inconsistent statistic for determining the extent of completed growth, leading to potentially biased inferences for research and applied purposes.