Cole T J
MRC Dunn Nutrition Unit, Cambridge, UK.
Ann Hum Biol. 1989 Sep-Oct;16(5):407-19. doi: 10.1080/03014468900000532.
The American National Center for Health Statistics (NCHS) and Dutch national height growth standards, unlike previous standards, allowed for skewness when fitting the centile curves. The LMS method (Cole 1988) is a way of summarizing growth standards which monitors the changing skewness of the distribution during childhood. It does so by calculating the Box-Cox power needed to transform the data to normality at each age, and displaying the results as a smooth curve of power plotted against age. Applying the LMS method to the NCHS and Dutch standards highlights the nature of the changing skewness of height during childhood, and particularly during puberty. Taken in conjunction with curves for the mean and coefficient of variation of height plotted against age, the power curve allows the original centiles to be reconstructed to high accuracy. This provides further evidence that the LMS method is a powerful and compact technique for deriving and presenting growth standards.
美国国家卫生统计中心（NCHS）和荷兰国家身高增长标准与以往标准不同，在拟合百分位数曲线时考虑了偏度。LMS方法（科尔，1988年）是一种总结生长标准的方法，它监测儿童期分布偏度的变化。该方法通过计算将各年龄数据转换为正态分布所需的Box-Cox幂，并将结果显示为幂随年龄变化的平滑曲线。将LMS方法应用于NCHS和荷兰标准，突出了儿童期尤其是青春期身高偏度变化的性质。结合身高均值和变异系数随年龄变化的曲线，幂曲线能够高精度地重建原始百分位数。这进一步证明了LMS方法是推导和呈现生长标准的一种强大而简洁的技术。
