Suppose we test a hypothesis that the means of a variable differ among four populations and then reject the null hypothesis. Does this assure us that each of the four means is significantly different from the other three? Explain.
I don't think it does but not 100%.