SECTION 14.2 Summary

• The ANOVA test does not say which of the group means differ. It is, therefore, usual to add comparisons among the means to basic ANOVA.
• Specific questions formulated before examination of the data can be expressed as contrasts. Tests and confidence intervals for contrasts provide answers to these questions.
• If no specific questions are formulated before examination of the data and if the null hypothesis of equality of population means is rejected, multiple-comparisons methods are used to assess the statistical significance of the differences between pairs of means. These methods are less powerful than contrasts, so use contrasts whenever a study is designed to answer specific questions.