Abstract
A computer simulation experiment was performed to examine the reliability and repeatability of the classification of the evenness variation Ⅱ test for raw silk in China. As the applicable grade was low, the sampling error was large for the standard deviation and small for the coefficient of variation. The appearance of the applicable grade was 46% for the same grade, 48% for differences in grade 1 and 6% for differences in grade 2. The probability of a defference in grade was 54%.
For repeated testing of the same lot, the differences in applicable grade were 34% for the same grade, 51% for differences in grade 1, 13% for differences in grade 2 and 1% for differences in grade 3. There was a 66% probability of there being a defference when grading twice. There was a marked difference between the experimental and calculated values. (Email : kouzu_ub@nifty.com)
