Abstract
To determine the coefficient of variation of certain characteristics in a yarn, such as the weight in constant short lengths, is a long and tedious task when using the standard root-mean-square method of computing the standard deviation. However, this standard deviation can be quite accurately estimated by any one of three short-cut methods described herein—the mean deviation method, the mean range method, and the probit method. The mean itself, by which this standard deviation is divided to get the coefficient of variation, is obtained readily by weighing all the samples together. The standard errors of the coefficient of variation estimated by these short-cut methods are derived, and the efficiencies of the methods are compared with that of the direct root- mean-square deviation method, showing the corresponding sample sizes for the same efficiency.