1952 年 16 巻 7 号 p. 412-415
In the factorial ore analysis of local research offices, I have stochastically considered the random, personal and sampling error, and from the results of the experiment the following conclusion was obtained: the personal error was of no significance when the level of significance was 0.05, and the random also was precise.
But the sampling precision was so bad that the analysis error happened to sample couldn’t be neglected.