Post Hoc and Planned Comparisons

  1. Introduction
    1. The ambiguity of a significant F and the purpose of post hoc comparisons
    2. Two features of planned comparisons
    3. The relative popularity of planned and post hoc comparisons
  2. Post Hoc Comparisons
    1. Definition and purpose
    2. Terminology
      1. Synonyms for "post hoc"
      2. "Contrast" and "comparison"
      3. The "omnibus" F-test
      4. "Pairwise" and "nonpairwise" comparisons
    3. Test procedures frequently used in post hoc analyses
      1. The 6 tests used most often by applied researchers
        1. Fisher's LSD
        2. Duncan's multiple range test
        3. Newman-Keuls test
        4. Bonferroni
        5. Tukey's test
        6. Scheffe's test
      2. "Liberal" vs. "conservative" procedures
      3. Comparing other groups vs. a control condition
        1. Dunnett's test
        2. Tamhane's test
      4. The null hypotheses of a post hoc investigation
      5. Presentation of results
        1. Results summarized in a passage of text
        2. Three different ways to summarize results in a table or figure
          1. Attaching letters to means
          2. Notes using "<" and ">" symbols (and maybe commas as well)
          3. Lines drawn above bars in a bargraph
    4. Planned Comparisons
      1. How these comparisons can be examined directly without reference to the omnibus F
      2. Test procedures used to make planned comparisons
      3. Two advantages of planned comparisons
      1. Terminology
        1. A priori
        2. One degree-of-freedom F-test
        3. Orthogonal
      2. Assumptions
      3. The researcher's choice of test procedure
      4. Statistical significance vs. practical significance
      5. Other test procedures


Copyright © 2012

Schuyler W. Huck
All rights reserved.

| Book Info | Author Info |

Site URL: www.readingstats.com

Top | Site Map
Site Design: John W. Taylor V