Marshall J D, Hazlett C B, Spady D W, Quinney H A
Department of Physical Education and Sport Studies, University of Alberta, Canada.
Am J Clin Nutr. 1990 Jan;51(1):22-8. doi: 10.1093/ajcn/51.1.22.
A comparative study of obesity measurements was undertaken with 533 male and female subjects, aged 11.8-15.9. Six adiposity measures (three skinfold indices, three height-weight indices) were derived from measures of height, weight, and eight skinfold thickness measurements. A principal components analysis of these adiposity measures resulted in a unifactorial solution accounting for 85.6% of the total variance. A cross-tabulation analysis with the derived factor scores and a criterion visual inspection rating supported the interpretation that the underlying construct of the factor was adiposity, and that a factor score of greater than 1.5 SD above the mean was a suitable standard for labeling obesity. Utilizing this dichotomy of factor scores as a standard, the differential diagnostic capabilities of four adiposity scales commonly used in identifying obesity was undertaken. Pursuit of this methodology, with the use of additional measures and larger sample sizes, is recommended to ensure the validation of an obesity measure.
对533名年龄在11.8至15.9岁之间的男性和女性受试者进行了肥胖测量的比较研究。从身高、体重测量以及八项皮褶厚度测量中得出了六种肥胖指标（三项皮褶指数、三项身高体重指数）。对这些肥胖指标进行主成分分析，得到了一个单因素解，占总方差的85.6%。将导出的因子得分与标准视觉检查评分进行交叉制表分析，支持了这样的解释：该因子的潜在结构是肥胖，且因子得分高于平均值1.5个标准差是标记肥胖的合适标准。以这种因子得分的二分法为标准，对常用于识别肥胖的四种肥胖量表的鉴别诊断能力进行了研究。建议采用这种方法，并使用额外的测量方法和更大的样本量，以确保肥胖测量方法的有效性。
