Select the group comparisons for One-Way ANOVA (Comparisons tab)

On the Comparisons tab of the One-Way ANOVA dialog box, select the comparison procedures to include in the analysis. The comparison procedures allow you to test the mean difference between pairs of groups. If you do not need to compare all groups, consider using a method that compares specific groups because these methods are more powerful. Select the comparison procedure based on the comparisons that are important for your situation.

Compare all groups

Tukey (family error rate)
Compares all pairs of groups, while controlling the simultaneous confidence level.
For example, a plastics company chooses Tukey's method to compare the strengths of five types of plastic, while maintaining a 95% simultaneous confidence level.
Fisher (individual error rate)
Compares all pairs of groups, while controlling the individual error rate. Fisher's method is less common than Tukey's method because it does not control the simultaneous confidence level, which often decreases to unacceptable levels.
For example, a plastics company chooses Fisher's method to compare the strengths of five types of plastic, while maintaining a 95% individual confidence level. After the analysis, the output indicates that the simultaneous confidence level is 71.73%, which is too low.

Compare to a defined group

Dunnett (to control group)
Compares the treatment groups to a control group, while controlling the simultaneous confidence level. For example, a plastics company chooses Dunnett's method to compare the strengths of four new plastic formulations to the formulation that the manufacturer currently uses.
From the drop-down list, select the control group, as follows:
  • If you chose Responses are in one column for all factor levels on the Data tab, select the value that identifies the control group.
  • If you chose Responses are in a separate column for each factor level on the Data tab, select the column that contains the control group.
Hsu MCB (to best group)
Identifies the groups that are the best. Hsu's MCB compares each group to the group that has either the largest or the smallest mean, while controlling the simultaneous confidence level. For example, a plastics company chooses Hsu's MCB to determine which of five plastic types is the strongest. For more information, go to Hsu's MCB.
From the drop-down list, select one of the following:
  • Largest mean is best: To compare each group to the group that has the largest mean.
  • Smallest mean is best: To compare each group to the group that has the smallest mean.
By using this site you agree to the use of cookies for analytics and personalized content.  Read our policy