Non-parametric tests Flashcards
What assumptions are common to hypothesis tests?
Randomness, independence, sample size.
What is randomness?
The sample is collected at random.
What is independence?
Each row is independent of the others.
What does sample size refer to?
If the sample is big enough.
Other assumption to consider for the data is…
To be normally distributed.
What assumption is common to test?
The sample size, and depends on the type of test.
Two-sample t-test size needed:
30 or higher.
One sample proportion test size needed:
10 or higher.
Chi-squared independence test size needed (across groups within):
5 or higher.
ANOVA test size needed:
30 or higher.
The Wilcoxon signed-rank test works well when?
The assumptions of a paired t-test aren’t met.
The Wilcoxon-Mann-Whitney test is like a?
Non-parametric t-test.
Useful when you can’t satisfy the assumptions for a parametric
test comparing two means.
The Kruskal-Wallis test is like a?
Non-parametric version of an ANOVA test, comparing the means across multiple groups.
When to consider to use a non-parametric test?
When the assumptions are not met (at least one).
