24小时热门版块排行榜

返回列表

当前只显示满足指定条件的回帖，点击这里查看本话题的所有回帖

tobebetter_

铜虫 (正式写手)

应助: 1 (幼儿园)
金币: 59.3
散金: 42
帖子: 308
在线: 54.1小时
虫号: 3636939
注册: 2015-01-06

[求助] 请问谁比较擅长统计相关的知识，什么是correlation t-test？有具体问题已有2人参与

我做的一个工作是模型模拟值与地面实测值进行比较。然后我有计算模拟值和实测值之间的相关性，就是最常用的皮尔逊相关系数，算出来的相关性的p-value<0.001.因此我认为我计算的相关性是有意义的。但是审稿人给我提了以下的问题，感觉涉及到了不少统计方面的知识，该如何理解？特别是他说到的"effective sample size autocorrelation t-test"怎么算，有现有软件或者程序可以算吗？真心求助，谢谢了！
"My single remaining concern has to do with the statistical significance tests I requested in my previous review. The authors have provided tests of significance, but based on the reported p-values and their own statement about sample size, I am concerned that the tests did not account for temporal autocorrelation. This would lead to an artificially high estimate of significance. A simple adjustment can be performed by using an effective sample size correction (an internet search for "effective sample size autocorrelation t-test" turns up several example of how this is done). "

回复此楼