我做的实验有四个实验条件,在进行repeated measure之后需要进行post hoc的比较。可是spss中不能直接作,下面是在网上搜到的解决方法: http://www.uvm.edu/~dhowell/StatPages/More_Stuff/RepMeasMultComp/RepMeasMultComp.html There are a number of reasons why standard software is not set up to run these comparisons easily. I suspect that the major reason is that unrestrained use of such procedures is generally unwise. Most people know that there are important assumptions behind repeated measures analysis of variance, most importantly the assumption of sphericity. Most people also know that there are procedures, such as the Greenhouse and Geisser and the Huynh and Feldt corrections, that allow us to deal with violations of sphericity. However many people do not know that those correction approaches become problematic when we deal with multiple comparisons, especially if we use an overall error term. The problem is that a correction factor computed on the full set of data does not apply well to tests based on only part of the data, so although th