[Updated Dec 2004 to correctly deal with repeated measures ANOVA.]
After two-way (or other) Analysis of Variance (ANOVA), you often wish to perform post tests to compare individual pairs of groups. Some programs don't perform post tests at all. Others, like GraphPad Prism, perform post tests for commonly-used experimental designs, but not for every experimental design. This calculator performs any post tests between pairs of cells that you select. But note that this calculator does not perform the ANOVA calculations -- you must use another program for that. Learn more about
(with example) and
how the calculations are performed
What about repeated measures ANOVA? If your post test compares the same set of subjects at two different times, then follow the instructions below. If your post test compare two sets of subjects at one time, then the
situation is trickier
. For that post test, the fact that you also measured the same subjects at other times is irrelevant. So the MSresidual reported in the ANOVA table is not the right value to use for the post test. Instead, rerun the analysis as an ordinary two-way ANOVA and enter that MSresidual (and corresonding DF value) in the form below. Before spending much time on these post tests,
consider whether two-way repeated-measures ANOVA is really the best analysis
If you are running post tests at every dose along a dose-response curve, or every time point along a kinetic curve, think hard about what you really want to know.
I don't think these post tests are very useful.