不放回抽样的样本均值方差是怎样推导出来的，是什么原因导致它与放回抽样的样本均值方差不一样的？
with replacement your obs are indep, without they are not, you will need a finite population correction, which is always less than one. This means that

when you sample without replacement, estimation of the population mean is a little

more precise.
谢谢啦！

看来我还得补习下英语。
还有人能具体的给讲一讲么？

最好能用中文。