OUTLINE FOR CHAPTER
18
Statistical Tests on Ranks (Nonparametric
Tests)
 Introduction
 Qualitative variables and nominal data
 The simplest kind of quantitative data
 The five most popular nonparametric tests
 Obtaining Ranked Data
 Having people(i.e., "judges") rank things
 Arranging people, animals, or things in rank order
 Converting raw scores into ranks
 Reasons for Converting Scores on a Continuous Variable
into Ranks
 Assumptions underlying test procedures
 Sample size
 The data's level of measurement
 Five Popular Test Procedures Used With Ranks
 The median test
 The appropriate "setting"
 The null hypothesis in the median test
 Moving from the raw data to a calculated value
 The meaning of a rejected null hypothesis
 The MannWhitney U test
 The appropriate "setting"
 Moving from the raw data to a calculated value
 The meaning of a rejected null hypothesis
 The "decision rule" for comparing the calculated and critical
values
 The KruskalWallis H test
 The appropriate "setting" and similarity to a oneway ANOVA
 Moving from the raw data to a calculated value
 The meaning of a rejected null hypothesis
 The "decision rule" for comparing the calculated and critical
values
 Post hoc investigations
 The Wilcoxon matchedpairs signedranks test
 The appropriate "setting"
 Moving from the raw data to a calculated value
 The meaning of a rejected null hypothesis
 Friedman's twoway analysis of variance of ranks
 The appropriate "setting" and similarity to a 1way repeatedmeasures
ANOVA
 Moving from the raw data to a calculated value
 The meaning of a rejected null hypothesis
 Related Issues
 Largesample versions of the tests on ranks
 The notion of a "largesample approximation"
 The use of z and chi square in tests on ranks
 The needed sample sizes for largesample approximations
 Ties
 How ties occur
 Three ways to deal with ties
 The relative power of nonparametric tests
 The meaning of "relative power"
 Conditions in which parametric tests have more power than
nonparametric tests and vice versa
 A Few Final Comments
 The quality of the research questions
 The assumptions of random samples and independence
 The term "distributionfree"
 Overlapping distributions
 Other nonparametric procedures
