Imagine that you are a statistical consultant and a researcher comes to you with a problem. The researcher has a random sample of data from each of six groups, and she wants to test whether the means of the six groups differ. In each of the following situations, recommend the best procedure.

a. The data are approximately normally distributed with equal standard deviations in all groups. Sample size is small.

b. The data are not normally distributed and do not have equal standard deviations in all groups. Sample size is small.

c. The data are not normally distributed. Groups have nearly equal standard deviations. Sample size is large and equal among groups.

d. The null hypothesis of no difference among group means was rejected. Now the researcher wants to find out which means are different. The data are approximately normally distributed with equal standard deviations in all groups, and the sample size is small.