Flashcards
Non-parametric tests
Cards in this deck (4)
- 1what are non-parametric tests? (Why are they good?)They do not involve population parametrics, such as requirements in ANOVA of distribution and independence.
Data can be measured on any scale - 2what are non-parametric tests? (Why are they good?)They do not involve population parametrics, such as requirements in ANOVA of distribution and independence.
Data can be measured on any scale - 3Mann-Whitney U1. non-parametric alternative to two-sample T-tests
2. uses ranks
3. Two-tailed Hypothesis e.g. H0= two predictors give same response
H1= two predictors give a different response - 4Kruskal-Wallis Test1. Can be used for 3+ groups
2. also uses rankings
avge. relative ratings should be about the same to accept H0.
3. Hypotheses: H0= Distributions of all the populations are the same. H1= Distributions of all populations are not the same.
BONUS= entire distributions compared
Information
- Description
- Flashcards on Non-parametric tests, created by marielouise.k on 16/05/2013.
- Created by
- marielouise.k
- Views
- 32
- Rating