non parametric tests Flashcards
when do we use non parametric tests?
when assumptions are not met for parametric
non parametric test for independent t test?
Mann Whitney U
non parametric tests for related t test?
Wilcoxon signed-ranks
non parametric test for Pearson’s product-moment correlation
spearman’s rho
when is non parametric tests appropriate to use
non normal distribution of data
outliers
small sample size
interval ratio or ordinal- DATA IS RANKED
Mann whitney U and Wilcoxon - effect sizes?
you have to calculate by hand
Z score divided by sqaure root of N
In spss for Mann whitney U tests when do you use which sig
when sample is below 40 you use exact sig 2
when we report results of Mann whiteny U tests what do we report
U SCORE
Z SCORE
P SCORE (sig)
EFFECT SIZE- u need to calculate- last sentence
MEDIANS (u have to ask spss to do this seperatly)
what do we report for wilcoxon
remember it gives u 2 tailed so if u did 1 tailed half the number to get you p value
Z SCORE
P SCORE (sig)
EFFECT SIZE (need to calculate- remembering N is total number of observations- often having to multiply N)
effect size for correlation
small-0.1-0.29
medium-0.3-0.49
large-0.5- 1
hypothesis in mann whiteny u
talk about the medians first- will they be equal or different?