统计分析的依据是样本。如果样本中混有系统误差,或者试验和数据处理中有差错引入了个别的反常数据,这个样本就不能代表母体,按此所作的一切统计分析都将失效。对于判别样本中的反常数据是否应予舍弃,数理统计中根据随机误差服从正态分布的规律,提供了一些科学的检验方法。此外,在生产管理中运用对反常数据的检验方法作为全面质量管理的主要 The basis of statistical analysis is the sample. If there is a systematic error in the sample, or when there are errors in the testing and data processing that introduce individual anomalies, the sample will not represent the parent and all statistical analysis done there will no longer be valid. For judging whether the abnormal data in the sample should be discarded or not, the mathematical statistics provide some scientific test methods according to the law that the random error obeys the normal distribution. In addition, the use of test methods for abnormal data in production management as the mainstay of total quality management