1. Multiple comparisons -- The example -- The null hypothesis -- Variance estimates or mean squares -- The F-ratio -- Decision rules and statistical significance -- Decision errors -- Selecting the most appropriate multiple-comparison test -- Selecting a multiple-comparison test and Type I errors -- Definitions and meanings of Type I errors in multiple-comparison tests -- Types of multiple comparisons -- Test statistics -- 2. Priori comparisons -- Orthogonal comparisons -- Characteristics of orthogonal comparisons -- Decision rule for evaluating orthogonal comparisons -- Strategies for developing sets of orthogonal comparisons -- Protection levels for orthogonal comparisons -- Orthogonal polynomials -- Nonorthogonal comparisons -- Summary and recommendations -- 3. Post hoc comparisons: pairwise methods -- Rationale for range tests -- Sampling distribution of ranges -- Tukey's honestly significant difference test (HSD) -- Newman-Keuls test -- Tukey's wholly significant difference test (WSD) -- The Duncan Test -- Lease significant difference test (LSD) -- Pairwise comparisons and Type I error -- Comparisons of pairwise methods -- Dunnett's test -- 4. Post Hoc comparisons: The Scheffé Test -- Controls for reducing Type I errors in post hoc comparisons -- Conducting the Scheffé test -- Characteristics of the Scheffé test -- 5. Multiple comparisons in factorial designs -- Familywise Type I error -- A nonquantitative approach for examining interactions -- Simple Main effects -- Multiple comparisons on interactions -- Range tests and interactions -- Planned comparisons and interaction effects -- 6. Additional considerations -- Unequal sample sizes -- Violations of assumptions -- Concluding remarks
Summary
'Multiple Comparisons' demonstrates the most important methods of investigating differences between levels of an independent variable within an experimental design. The authors review the analysis of variance and hypothesis testing and describe the dimensions on which multiple comparisons vary