中文摘要：

在气候诊断研究和短期气候预测中广泛使用统计方法，例如相关分析和合成分析。本文对国内外文献中对统计检验显著性水平的各种描述进行了评论，指出了错误。指出了描述统计检验的结果时应该注意的问题。认为，对相关系数和合成分析等统计检验的描述，不应该称为“95％的信度”，“95％的显著性水平”，而应该称为“5％的信度（显著性水平）”，或者“95％的置信水平”。还指出，统计检验或者显著性检验不应该被改称为信度检验，并分析了错误与产生混乱的原因。

英文摘要：

Statistics method, for example, the correlation analysis and composite analysis, has already been used extensively in the climate diagnosis research and short-term climate forecasting. This article aimed at the comment on various descriptions of the significance level of the statistics test in the domestic and international articles and submit the problem that the description of statistics test should take into account, and point out the bug of various descriptions. The conclusions are that description of the significance level of the statistics test should be called as 5 % significance level or 95 % confidence level, rather than as 95% significance level. It also pointed out that the statistics test or significance test shouldn＇t be called the significance level test. The cause of the bug and the confusion for description of the significance level of the statistics test are analyzed.