PSYC 3000 Lecture Notes - Lecture 29: Wilcoxon Signed-Rank Test, Nonparametric Statistics, Parametric Statistics

• School: Carleton University
• Department: Psychology
• Course: PSYC 3000
• Professor: Bruce Hutcheon

Document Summary

Analogues of anova = tests you run when you can"t run anovas (when the condiions of anova aren"t met) For anova -----> dv = interval or beter. B/c this dv is interval at least, that means we can calculate variances, means, etc. ----> can"t do this without at least interval data. Behind anova, we assume there"s a populaion that has a normal distribuion with a given mean, variance, skew, kurtosis (i. e. populaion parameters) -----> this cannot be stated if the dv isn"t at least interval. ** this scheme involves parametric staisics (t-tests and z-tests are also parametrict) Called non-parametric b/c the assumpion of the populaion behind the samples do not have parameters to describe them // we don"t describe the distribuion in the populaion. Kruskal-wallis test -----> when the dv is ordinal (not as heavily structured as interval data) We can keep track of the ranks; we can run mean rank tests.