OUTLINE FOR CHAPTER 18

Statistical Tests on Ranks (Nonparametric Tests)

  1. Introduction
    1. Qualitative variables and nominal data
    2. The simplest kind of quantitative data
    3. The five most popular nonparametric tests
  2. Obtaining Ranked Data
    1. Having people(i.e., "judges") rank things
    2. Arranging people, animals, or things in rank order
    3. Converting raw scores into ranks
  3. Reasons for Converting Scores on a Continuous Variable into Ranks
    1. Assumptions underlying test procedures
    2. Sample size
    3. The data's level of measurement
  4. Five Popular Test Procedures Used With Ranks
    1. The median test
      1. The appropriate "setting"
      2. The null hypothesis in the median test
      3. Moving from the raw data to a calculated value
      4. The meaning of a rejected null hypothesis
    2. The Mann-Whitney U test
      1. The appropriate "setting"
      2. Moving from the raw data to a calculated value
      3. The meaning of a rejected null hypothesis
      4. The "decision rule" for comparing the calculated and critical values
    3. The Kruskal-Wallis H test
      1. The appropriate "setting" and similarity to a one-way ANOVA
      2. Moving from the raw data to a calculated value
      3. The meaning of a rejected null hypothesis
      4. The "decision rule" for comparing the calculated and critical values
      5. Post hoc investigations
    4. The Wilcoxon matched-pairs signed-ranks test
      1. The appropriate "setting"
      2. Moving from the raw data to a calculated value
      3. The meaning of a rejected null hypothesis
    5. Friedman's two-way analysis of variance of ranks
      1. The appropriate "setting" and similarity to a 1-way repeated-measures ANOVA
      2. Moving from the raw data to a calculated value
      3. The meaning of a rejected null hypothesis
  5. Related Issues
    1. Large-sample versions of the tests on ranks
      1. The notion of a "large-sample approximation"
      2. The use of z and chi square in tests on ranks
      3. The needed sample sizes for large-sample approximations
    2. Ties
      1. How ties occur
      2. Three ways to deal with ties
    3. The relative power of nonparametric tests
      1. The meaning of "relative power"
      2. Conditions in which parametric tests have more power than nonparametric tests and vice versa
  6. A Few Final Comments
    1. The quality of the research questions
    2. The assumptions of random samples and independence
    3. The term "distribution-free"
    4. Overlapping distributions
    5. Other nonparametric 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