Ward R, Schlenker J, Anderson G S
School of Kinesiology, Simon Fraser University, Burnaby, British Columbia V5A 1S6, Canada.
Am J Phys Anthropol. 2001 Nov;116(3):246-50. doi: 10.1002/ajpa.1120.
The present paper demonstrates the ease of use of method I by Preece and Baine ([1978] Ann Hum Biol 5:1-24) in generating smoothed growth curves for both height and weight. Using the National Center for Health Statistics (NCHS) growth curve data, smoothed curves were developed and compared to those produced using the least-squares-cubic-spline method. Based on the lower sum of squares and better fit of shape as indicated by residual examination, it was concluded that the method I curve fitting procedure by Preece and Baine ([1978] Ann Hum Biol 5:1-24) fit centile growth curves for height and weight in 2-18-year-old male and female children as well as, if not better than, the least-squares-cubic-spline method used in developing the 1979 NCHS growth curves. Further, as this paper demonstrates, smoothed curves can be generated on a desktop computer using readily available software (the SOLVER function within Microsoft EXCEL).
本文展示了普里斯和贝恩([1978]《人类生物学杂志》5:1 - 24)的方法I在生成身高和体重的平滑生长曲线方面的易用性。利用美国国家卫生统计中心(NCHS)的生长曲线数据,绘制了平滑曲线,并与使用最小二乘三次样条法生成的曲线进行了比较。基于较低的平方和以及残差检验所表明的更好的形状拟合,得出结论:普里斯和贝恩([1978]《人类生物学杂志》5:1 - 24)的方法I曲线拟合程序在拟合2至18岁男童和女童的身高和体重百分位数生长曲线方面,即便不比用于制定1979年NCHS生长曲线的最小二乘三次样条法更好,至少也与之相当。此外,正如本文所展示的,使用现成软件(微软Excel中的规划求解功能)可以在台式计算机上生成平滑曲线。