Harris E F, Weinstein S, Weinstein L, Poole A E
Ann Hum Biol. 1980 May-Jun;7(3):225-34. doi: 10.1080/03014468000004271.
Three methods of predicting a child's adult height are in common use: Bayley and Pinneau (1952), Tanner et al. (1975) and Roche et al. (1975). The relative accuracies of these methods have been assessed using growth records (22 male, 24 female) from the Child Research Council, Denver, Colorado. A series of six ages from five years to mid-adolescence was examined. Testing all methods and ages within each sex with a multivariate design yielded significant differences in the methods' accuracies, but inspection of the data by age disclosed that the major discrimination occurred during adolescence, not childhood. Results also indicated that the method of skeletal age assessment (either Greulich--Pyle or TW2) is more critical to accurate height prediction than is the choice of prediction method per se. This is because of inter-population differences in the rate and pattern of progress towards maturity and thus indicates the need to compare the children under examination to the most appropriate population standards available.
贝利和皮诺（1952年）、坦纳等人（1975年）以及罗氏等人（1975年）提出的方法。已使用科罗拉多州丹佛市儿童研究委员会的生长记录（22名男性、24名女性）对这些方法的相对准确性进行了评估。研究了从五岁到青春期中期的六个年龄段。采用多变量设计对每种性别中的所有方法和年龄段进行测试，结果显示这些方法的准确性存在显著差异，但按年龄检查数据发现，主要的差异出现在青春期，而非儿童期。结果还表明，骨骼年龄评估方法（格吕利希-派尔法或TW2法）对准确预测身高比预测方法本身的选择更为关键。这是因为不同人群在走向成熟的速度和模式上存在差异，因此表明有必要将受检儿童与现有的最合适人群标准进行比较。
