1
Q
Non-parametric tests
A
Does NOT have
* Normality (Skewed Data)
* Variacance (Homogeneity is violated)
Examines the difference between groups or variable (essentially a t-test)
2
Q
Two criteria for choosing non-parametric tests
A
- Normality and homogeneity of variance are not satisfied
- Data measured on the nominal or ordinal scale
3
Q
Non parametric is less ____ because it is a rank and not interval/ratio data
A
powerful
4
Q
Disadvantages: non-parametrics
A
- Can’t do complex designs (T-Test or One Way ANOVA)
- Less powerful and need a greater sample size (10 subjects for every 6 parametric)
- Still need to have a randomization procedure to assign groups
5
Q
Test for 2 independent groups is called
A
Mann-Whitney U
6
Q
Two dependent Samples Test is called
A
Wilcoxon signed ranks test
7
Q
K Related Independent Test is called
A
Kruskal-Wallis Oneway Analysis
8
Q
K Related Dependent Variables
A
- Friedman Oneway Analysis of Variance
9
Q
What central tendency do we use for non parametrics?
A
median
