气象中使用统计检验的几个问题

Some Problems on Using Statistical Test in Meteorology

针对目前在气象和气候要素变化分析中,对相关系数和合成分析使用统计检验存在的若干问题进行评论。认为对相关系数等统计检验,不应该称为“信度检验”,其检验水平也不应该称为“95%置信度”,而应该称为“显著性检验”,“显著性水平为5%”。

The comment on statistical test of correlation coefficients and others are given. The statistical test is used to apply to some analysis of climate variation of meteorological elements. The conclusions are that it should be call as “significant test” and “significant level of 5%”, rather than as “confidence test” and “confidence level of 95%”.