传统的方差分析方法,在单一衡量指标的一维空间中,是人们检测处理真正差异的有效手段。但在不具有可比性的多个衡量指标的多维空间中,方差分析则难于采用。而这种现象在科研和生产中是大量存在的。如保花保果调节剂混用效果的好与差,不仅要从座果率、产量差异上判别,而且还应对 The traditional method of analysis of variance, in a one-dimensional space of a single metric, is an effective way for people to detect and deal with real differences. However, analysis of variance is difficult to apply in multidimensional space without multiple measures of comparability. However, this kind of phenomenon exists in a large amount in scientific research and production. Such as flower and fruit conditioner mixture effect is good and bad, not only from the fruit rate, yield differences on the judge, but also to deal with